diff --git a/.gitignore b/.gitignore
index fbab1b2097fad956dceb05769613cb82053b27e8..63bd843b606911d303ce1476fc002c009b8ebc9a 100644
--- a/.gitignore
+++ b/.gitignore
@@ -8,8 +8,10 @@ run/
_skbuild/
dist/
*.egg-info/
-__pycache__/
+**/__pycache__/
.venv/
+pythonbindings/pyfluids/bindings*
+pythonbindings/pymuparser/bindings*
# IDE
.vscode/
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 22cc5bdf03e3870b94327bf207d9ce163a321013..e6ef4c9a0584c55adc9f745bcfdca4d1bbb08fbe 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -126,13 +126,14 @@ gcc_9_python:
paths:
- build/
- dist/
+ - _skbuild/
before_script:
- export CCACHE_BASEDIR=$CI_PROJECT_DIR
- export CCACHE_DIR=$CI_PROJECT_DIR/cache
script:
- - python3 setup.py bdist_wheel build_ext --build-temp=build
+ - python3 setup.py bdist_wheel build_ext --build-temp=_skbuild -- -DBUILD_VF_CPU=ON -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DCMAKE_CUDA_COMPILER_LAUNCHER=ccache -DCMAKE_C_COMPILER_LAUNCHER=ccache
###############################################################################
## Container Upload ##
diff --git a/CMake/cmake_config_files/MULE.config.cmake b/CMake/cmake_config_files/MULE.config.cmake
index 02f61b7988c5b3af9cd58bc52e46b1b2edfe8aae..2afbce6cc257fa0b8ff4dd7de580cb50c01369f1 100644
--- a/CMake/cmake_config_files/MULE.config.cmake
+++ b/CMake/cmake_config_files/MULE.config.cmake
@@ -1 +1,4 @@
-SET(CMAKE_CUDA_ARCHITECTURES "75")
\ No newline at end of file
+SET(CMAKE_CUDA_ARCHITECTURES "75")
+
+list(APPEND USER_APPS "apps/gpu/LBM/ActuatorLine")
+list(APPEND USER_APPS "apps/gpu/LBM/SphereScaling")
diff --git a/CMakePresets.json b/CMakePresets.json
index 0f360fd303cdcad923b01d56df5c6d48ad62ca2c..c53482ec72109f1a672b97797763d027a6ec80bf 100644
--- a/CMakePresets.json
+++ b/CMakePresets.json
@@ -9,26 +9,29 @@
{
"name": "default",
"hidden": true,
- "binaryDir": "${sourceDir}/build/",
- "cacheVariables": {
- "BUILD_VF_UNIT_TESTS": "ON"
- }
+ "binaryDir": "${sourceDir}/build/"
},
{
- "name": "default_make",
- "inherits": "default",
+ "name": "msvc",
+ "hidden": true,
+ "generator": "Visual Studio 16 2019",
+ "architecture": "x64"
+ },
+ {
+ "name": "make",
"hidden": true,
"generator": "Unix Makefiles"
},
{
- "name": "default_msvc",
- "inherits": "default",
+ "name": "unit_tests",
"hidden": true,
- "generator": "Visual Studio 16 2019",
- "architecture": "x64"
+ "cacheVariables": {
+ "BUILD_VF_UNIT_TESTS": "ON"
+ }
},
{
- "name": "default_cpu",
+ "name": "cpu",
+ "inherits": "default",
"hidden": true,
"description": "CPU build of VirtualFluids",
"cacheVariables": {
@@ -37,7 +40,8 @@
}
},
{
- "name": "default_gpu",
+ "name": "gpu",
+ "inherits": "default",
"hidden": true,
"description": "GPU build of VirtualFluids",
"cacheVariables": {
@@ -46,92 +50,213 @@
}
},
{
- "name": "default_gpu_numerical_tests",
- "inherits": [
- "default_gpu"
- ],
+ "name": "debug",
"hidden": true,
- "description": "GPU numerical tests of VirtualFluids",
"cacheVariables": {
- "BUILD_VF_DOUBLE_ACCURACY": "ON",
- "BUILD_NUMERIC_TESTS": "ON"
+ "CMAKE_BUILD_TYPE": "Debug"
}
},
{
- "name": "default_all",
+ "name": "release",
"hidden": true,
- "description": "All build of VirtualFluids",
- "inherits": [
- "default_cpu",
- "default_gpu"
- ],
"cacheVariables": {
- "BUILD_VF_DOUBLE_ACCURACY": "ON"
+ "CMAKE_BUILD_TYPE": "Release"
}
},
{
- "name": "cpu_make",
- "inherits": [
- "default_make",
- "default_cpu"
- ],
- "displayName": "cpu make configuration"
- },
- {
- "name": "cpu_msvc",
- "inherits": [
- "default_msvc",
- "default_cpu"
- ],
- "displayName": "cpu msvc configuration"
+ "name": "min_size_rel",
+ "hidden": true,
+ "cacheVariables": {
+ "CMAKE_BUILD_TYPE": "MinSizeRel"
+ }
},
{
- "name": "gpu_make",
- "inherits": [
- "default_make",
- "default_gpu"
- ],
- "displayName": "gpu make configuration"
+ "name": "rel_with_deb_info",
+ "hidden": true,
+ "cacheVariables": {
+ "CMAKE_BUILD_TYPE": "RelWithDebInfo"
+ }
},
{
- "name": "gpu_msvc",
+ "name": "gpu_numerical_tests",
"inherits": [
- "default_msvc",
- "default_gpu"
+ "gpu",
+ "unit_tests"
],
- "displayName": "gpu msvc configuration"
+ "hidden": true,
+ "description": "GPU numerical tests of VirtualFluids",
+ "cacheVariables": {
+ "BUILD_VF_DOUBLE_ACCURACY": "ON",
+ "BUILD_NUMERIC_TESTS": "ON"
+ }
},
{
"name": "all_make",
"inherits": [
- "default_make",
- "default_all"
+ "cpu",
+ "gpu",
+ "unit_tests",
+ "make"
],
"displayName": "all make configuration"
},
{
"name": "all_msvc",
"inherits": [
- "default_msvc",
- "default_all"
+ "cpu",
+ "gpu",
+ "unit_tests",
+ "msvc"
],
"displayName": "all msvc configuration"
},
{
"name": "gpu_numerical_tests_make",
"inherits": [
- "default_make",
- "default_gpu_numerical_tests"
+ "gpu_numerical_tests",
+ "make"
],
"displayName": "gpu numerical tests make configuration"
},
{
"name": "gpu_numerical_tests_msvc",
"inherits": [
- "default_msvc",
- "default_gpu_numerical_tests"
+ "msvc",
+ "gpu_numerical_tests"
],
"displayName": "gpu numerical tests msvc configuration"
+ },
+ {
+ "name": "debug_make_gpu",
+ "displayName": "Debug GPU Make",
+ "inherits": [
+ "gpu",
+ "make",
+ "debug"
+ ]
+ },
+ {
+ "name": "release_make_gpu",
+ "displayName": "Release GPU Make",
+ "inherits": [
+ "gpu",
+ "make",
+ "release"
+ ]
+ },
+ {
+ "name": "min_size_rel_make_gpu",
+ "displayName": "MinSizeRel GPU Make",
+ "inherits": [
+ "gpu",
+ "make",
+ "min_size_rel"
+ ]
+ },
+ {
+ "name": "rel_with_deb_info_make_gpu",
+ "displayName": "RelWithDebInfo GPU Make",
+ "inherits": [
+ "gpu",
+ "make",
+ "rel_with_deb_info"
+ ]
+ },
+ {
+ "name": "debug_msvc_gpu",
+ "displayName": "Debug GPU MSVC",
+ "inherits": [
+ "gpu",
+ "msvc",
+ "debug"
+ ]
+ },
+ {
+ "name": "release_msvc_gpu",
+ "displayName": "Release GPU MSVC",
+ "inherits": [
+ "gpu",
+ "msvc",
+ "release"
+ ]
+ },
+ {
+ "name": "min_size_rel_msvc_gpu",
+ "displayName": "MinSizeRel GPU MSVC",
+ "inherits": [
+ "gpu",
+ "msvc",
+ "min_size_rel"
+ ]
+ },
+ {
+ "name": "rel_with_deb_info_msvc_gpu",
+ "displayName": "RelWithDebInfo GPU MSVC",
+ "inherits": [
+ "gpu",
+ "msvc",
+ "rel_with_deb_info"
+ ]
+ }
+ ],
+ "buildPresets": [
+ {
+ "name": "Default",
+ "hidden": true,
+ "configurePreset": "default",
+ "jobs": 4
+ },
+ {
+ "name": "GPU",
+ "hidden": true,
+ "configurePreset": "gpu",
+ "targets": [
+ "ActuatorLine",
+ "DrivenCavity",
+ "BoundaryLayer"
+ ],
+ "inherits": [
+ "Default"
+ ]
+ },
+ {
+ "name": "Release",
+ "hidden": true,
+ "configurePreset": "release"
+ },
+ {
+ "name": "Debug_Make_GPU",
+ "displayName": "Debug",
+ "description": "Compile GPU version with debug information",
+ "configurePreset": "debug_make_gpu",
+ "inherits": [
+ "GPU"
+ ]
+ },
+ {
+ "name": "MinSizeRel_Make_GPU",
+ "displayName": "MinSizeRel",
+ "configurePreset": "min_size_rel_make_gpu",
+ "inherits": [
+ "GPU"
+ ]
+ },
+ {
+ "name": "RelWithDebInfo_GPU",
+ "displayName": "RelWithDebInfo",
+ "configurePreset": "rel_with_deb_info_make_gpu",
+ "inherits": [
+ "GPU"
+ ]
+ },
+ {
+ "name": "Release_GPU",
+ "description": "Build release version of GPU",
+ "displayName": "Release GPU",
+ "configurePreset": "release_make_gpu",
+ "inherits": [
+ "GPU"
+ ]
}
]
-}
+}
\ No newline at end of file
diff --git a/MANIFEST.in b/MANIFEST.in
new file mode 100644
index 0000000000000000000000000000000000000000..adafcf99560acd9da79aa060194df8263b6e77e0
--- /dev/null
+++ b/MANIFEST.in
@@ -0,0 +1 @@
+include pythonbindings/*/bindings*
\ No newline at end of file
diff --git a/Python/SlurmTests/poiseuille/settings.py b/Python/SlurmTests/poiseuille/settings.py
index 4b4a1e4e9cc7f6118a60c22a40c70b027e3ac4e2..a3cdc5dc8b627612c2d57a58db36c9fbaa72efac 100644
--- a/Python/SlurmTests/poiseuille/settings.py
+++ b/Python/SlurmTests/poiseuille/settings.py
@@ -1,25 +1,58 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file settings.py
+! \ingroup Poiseuille
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
import os
from acousticscaling import OneDirectionalAcousticScaling
-from pyfluids.cpu.kernel import LBMKernel, KernelType
-from pyfluids.cpu.parameters import RuntimeParameters, GridParameters, PhysicalParameters
+from pyfluids import cpu
-grid_params = GridParameters()
+grid_params = cpu.parameters.GridParameters()
grid_params.node_distance = 1
grid_params.number_of_nodes_per_direction = [1, 1, 16]
grid_params.blocks_per_direction = [1, 1, 4]
grid_params.periodic_boundary_in_x1 = True
grid_params.periodic_boundary_in_x2 = True
-physical_params = PhysicalParameters()
+physical_params = cpu.parameters.PhysicalParameters()
physical_params.lattice_viscosity = 1e-4
-runtime_params = RuntimeParameters()
+runtime_params = cpu.parameters.RuntimeParameters()
runtime_params.number_of_threads = int(os.environ["PYFLUIDS_NUM_THREADS"])
runtime_params.number_of_timesteps = 4_000_000
runtime_params.timestep_log_interval = 1_000_000
-kernel = LBMKernel(KernelType.CompressibleCumulantFourthOrderViscosity)
+kernel = cpu.kernel.LBMKernel(cpu.kernel.KernelType.CompressibleCumulantFourthOrderViscosity)
kernel.use_forcing = True
kernel.forcing_in_x1 = 5e-10
diff --git a/Python/acousticscaling.py b/Python/acousticscaling.py
index a664b8e924d648b680562b9aef11bee87b3562b1..7e71fed9fdd9f86415261ef4e22797021581f60c 100644
--- a/Python/acousticscaling.py
+++ b/Python/acousticscaling.py
@@ -1,22 +1,55 @@
-from pyfluids.cpu.kernel import LBMKernel
-from pyfluids.cpu.parameters import GridParameters, PhysicalParameters, RuntimeParameters
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file acousticscaling.py
+! \ingroup tests
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
+from pyfluids import cpu
class OneDirectionalAcousticScaling:
- def __init__(self, grid_parameters: GridParameters,
- physical_parameters: PhysicalParameters,
- runtime_parameters: RuntimeParameters,
- kernel: LBMKernel):
+ def __init__(self, grid_parameters: cpu.parameters.GridParameters,
+ physical_parameters: cpu.parameters.PhysicalParameters,
+ runtime_parameters: cpu.parameters.RuntimeParameters,
+ kernel: cpu.kernel.LBMKernel):
self._grid_params = grid_parameters
self._physical_params = physical_parameters
self._runtime_params = runtime_parameters
self._kernel = kernel
- def configuration_for_scale_level(self, level: int = 1) -> tuple[GridParameters,
- PhysicalParameters,
- RuntimeParameters,
- LBMKernel]:
+ def configuration_for_scale_level(self, level: int = 1) -> tuple[cpu.parameters.GridParameters,
+ cpu.parameters.PhysicalParameters,
+ cpu.parameters.RuntimeParameters,
+ cpu.kernel.LBMKernel]:
if level < 0:
raise ValueError("level must be >= 0")
@@ -27,8 +60,8 @@ class OneDirectionalAcousticScaling:
return grid_params, physical_params, runtime_params, kernel
- def clone_grid_params_for_level(self, level) -> GridParameters:
- grid_params = GridParameters()
+ def clone_grid_params_for_level(self, level) -> cpu.parameters.GridParameters:
+ grid_params = cpu.parameters.GridParameters()
grid_params.reference_direction_index = self._grid_params.reference_direction_index
grid_params.periodic_boundary_in_x1 = self._grid_params.periodic_boundary_in_x1
grid_params.periodic_boundary_in_x2 = self._grid_params.periodic_boundary_in_x2
@@ -51,7 +84,7 @@ class OneDirectionalAcousticScaling:
return grid_params
def clone_physical_parameters(self, level):
- physical_params = PhysicalParameters()
+ physical_params = cpu.parameters.PhysicalParameters()
physical_params.lattice_viscosity = self._physical_params.lattice_viscosity
if level > 0:
@@ -60,7 +93,7 @@ class OneDirectionalAcousticScaling:
return physical_params
def clone_runtime_params_for_level(self, level):
- runtime_params = RuntimeParameters()
+ runtime_params = cpu.parameters.RuntimeParameters()
runtime_params.number_of_timesteps = self._runtime_params.number_of_timesteps
runtime_params.number_of_threads = self._runtime_params.number_of_threads
runtime_params.timestep_log_interval = self._runtime_params.timestep_log_interval
@@ -71,7 +104,7 @@ class OneDirectionalAcousticScaling:
return runtime_params
def clone_kernel_for_level(self, level):
- kernel = LBMKernel(self._kernel.type)
+ kernel = cpu.kernel.LBMKernel(self._kernel.type)
kernel.use_forcing = self._kernel.use_forcing
kernel.forcing_in_x1 = self._kernel.forcing_in_x1
kernel.forcing_in_x2 = self._kernel.forcing_in_x2
diff --git a/Python/actuator_line/actuator_line.py b/Python/actuator_line/actuator_line.py
index 6e3c8608617df1267535984d53307dea9184c6ab..721af737ff6ef3340c3c2f6204aa6a7824cd1d2f 100644
--- a/Python/actuator_line/actuator_line.py
+++ b/Python/actuator_line/actuator_line.py
@@ -1,23 +1,48 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file actuator_line.py
+! \ingroup actuator_line
+! \author Henry Korb, Henrik Asmuth
+=======================================================================================
+"""
#%%
import numpy as np
from pathlib import Path
from mpi4py import MPI
-from pyfluids import basics, gpu, logger
+from pyfluids.bindings import basics, gpu, logger
#%%
-reference_diameter = 126
-
-length = np.array([29,6,6])*reference_diameter
-viscosity = 1.56e-5
-velocity = 9
-mach = 0.1
-nodes_per_diameter = 32
-
-sim_name = "ActuatorLine"
-config_file = Path(__file__).parent/Path("config.txt")
+sim_name = "ABL"
+config_file = Path(__file__).parent/"configActuatorLine.txt"
output_path = Path(__file__).parent/Path("output")
output_path.mkdir(exist_ok=True)
-t_out = 100.
-t_end = 500.
+
#%%
logger.Logger.initialize_logger()
@@ -25,87 +50,175 @@ basics.logger.Logger.add_stdout()
basics.logger.Logger.set_debug_level(basics.logger.Level.INFO_LOW)
basics.logger.Logger.time_stamp(basics.logger.TimeStamp.ENABLE)
basics.logger.Logger.enable_printed_rank_numbers(True)
-# %%
-comm = gpu.Communicator.get_instance()
#%%
grid_factory = gpu.grid_generator.GridFactory.make()
grid_builder = gpu.grid_generator.MultipleGridBuilder.make_shared(grid_factory)
+communicator = gpu.Communicator.get_instance()
-#%%
-dx = reference_diameter/nodes_per_diameter
-
-grid_builder.add_coarse_grid(0.0, 0.0, 0.0, *length, dx)
-grid_builder.set_periodic_boundary_condition(False, False, False)
-grid_builder.build_grids(basics.LbmOrGks.LBM, False)
-#%%
config = basics.ConfigurationFile()
config.load(str(config_file))
+
+para = gpu.Parameter(communicator.get_number_of_process(), communicator.get_pid(), config)
+bc_factory = gpu.BoundaryConditionFactory()
+
#%%
-para = gpu.Parameter(config, comm.get_number_of_process(), comm.get_pid())
+turbine_diameter = config.get_float_value("turbineDiameter", 126)
+boundary_layer_height = config.get_float_value("boundaryLayerHeight", 1000)
+z0 = config.get_float_value("z0", 0.1)
+u_star = config.get_float_value("u_star", 0.4)
+
+kappa = config.get_float_value("vonKarmanConstant", 0.4) # von Karman constant
+
+viscosity = config.get_float_value("viscosity", 1.56e-5)
+
+velocity = 0.5*u_star/kappa*np.log(boundary_layer_height/z0+1) #0.5 times max mean velocity at the top in m/s
+
+mach = config.get_float_value("Ma", 0.1)
+nodes_per_height = config.get_uint_value("nz", 64)
+
+
+turb_pos = np.array([3,3,3])*turbine_diameter
+epsilon = config.get_float_value("SmearingWidth", 5)
+density = config.get_float_value("Density", 1.225)
+level = 0
+n_blades = 3
+n_blade_nodes = config.get_int_value("NumberOfNodesPerAL", 32)
+
+read_precursor = config.get_bool_value("readPrecursor", False)
+
+if read_precursor:
+ nTReadPrecursor = config.get_int_value("nTimestepsReadPrecursor")
+ use_distributions = config.get_bool_value("useDistributions", False)
+ precursor_directory = config.get_string_value("precursorDirectory")
+
+# all in s
+t_start_out = config.get_float_value("tStartOut")
+t_out = config.get_float_value("tOut")
+t_end = config.get_float_value("tEnd") # total time of simulation
+t_start_averaging = config.get_float_value("tStartAveraging")
+t_start_tmp_averaging = config.get_float_value("tStartTmpAveraging")
+t_averaging = config.get_float_value("tAveraging")
+t_start_out_probe = config.get_float_value("tStartOutProbe")
+t_out_probe = config.get_float_value("tOutProbe")
+
+#%%
+length = np.array([6,4,1])*boundary_layer_height
+dx = boundary_layer_height/nodes_per_height
dt = dx * mach / (np.sqrt(3) * velocity)
-velocity_lb = velocity * dt / dx # LB units
-viscosity_lb = viscosity * dt / (dx * dx) # LB units
+velocity_ratio = dx/dt
+velocity_LB = velocity / velocity_ratio # LB units
+viscosity_LB = viscosity / (velocity_ratio * dx) # LB units
+pressure_gradient = u_star * u_star / boundary_layer_height
+pressure_gradient_LB = pressure_gradient * (dt*dt)/dx
+
+logger.vf_log_info(f"velocity [dx/dt] = {velocity_LB}")
+logger.vf_log_info(f"dt = {dt}")
+logger.vf_log_info(f"dx = {dx}")
+logger.vf_log_info(f"viscosity [10^8 dx^2/dt] = {viscosity_LB*1e8}")
+logger.vf_log_info(f"u* /(dx/dt) = {u_star*dt/dx}")
+logger.vf_log_info(f"dpdx = {pressure_gradient}")
+logger.vf_log_info(f"dpdx /(dx/dt^2) = {pressure_gradient_LB}")
+
#%%
-para.set_devices([0])
para.set_output_prefix(sim_name)
-para.set_output_path(str(output_path))
-para.set_f_name(para.get_output_path() + "/" + para.get_output_prefix())
para.set_print_files(True)
-para.set_max_level(1)
-#%%
-para.set_velocity(velocity_lb)
-para.set_viscosity(viscosity_lb)
+
+para.set_forcing(pressure_gradient_LB, 0, 0)
+para.set_velocity_LB(velocity_LB)
+para.set_viscosity_LB(viscosity_LB)
para.set_velocity_ratio(dx/dt)
para.set_viscosity_ratio(dx*dx/dt)
-para.set_main_kernel("TurbulentViscosityCumulantK17CompChim")
-para.set_use_AMD(True)
-para.set_SGS_constant(0.083)
+para.set_density_ratio(1.0)
-def init_func(coord_x, coord_y, coord_z):
- return [0.0, velocity_lb, 0.0, 0.0]
+para.set_main_kernel("CumulantK17")
-para.set_initial_condition(init_func)
-para.set_t_out(int(t_out/dt))
-para.set_t_end(int(t_end/dt))
+para.set_timestep_start_out(int(t_start_out/dt))
+para.set_timestep_out(int(t_out/dt))
+para.set_timestep_end(int(t_end/dt))
para.set_is_body_force(True)
-
#%%
-grid_builder.set_velocity_boundary_condition(gpu.SideType.MX, velocity_lb, 0.0, 0.0)
-
-grid_builder.set_velocity_boundary_condition(gpu.SideType.MY, velocity_lb, 0.0, 0.0)
-grid_builder.set_velocity_boundary_condition(gpu.SideType.PY, velocity_lb, 0.0, 0.0)
-
-grid_builder.set_velocity_boundary_condition(gpu.SideType.MZ, velocity_lb, 0.0, 0.0)
-grid_builder.set_velocity_boundary_condition(gpu.SideType.PZ, velocity_lb, 0.0, 0.0)
+tm_factory = gpu.TurbulenceModelFactory(para)
+tm_factory.read_config_file(config)
+#%%
+grid_scaling_factory = gpu.GridScalingFactory()
+grid_scaling_factory.set_scaling_factory(gpu.GridScaling.ScaleCompressible)
-grid_builder.set_pressure_boundary_condition(gpu.SideType.PX, 0.0)
+grid_builder.add_coarse_grid(0.0, 0.0, 0.0, *length, dx)
+grid_builder.set_periodic_boundary_condition(not read_precursor, True, False)
+grid_builder.build_grids(basics.LbmOrGks.LBM, False)
+sampling_offset = 2
+if read_precursor:
+ precursor = gpu.create_file_collection(precursor_directory + "/precursor", gpu.FileType.VTK)
+ grid_builder.set_precursor_boundary_condition(gpu.SideType.MX, precursor, nTReadPrecursor, 0, 0, 0)
+
+grid_builder.set_stress_boundary_condition(gpu.SideType.MZ, 0, 0, 1, sampling_offset, z0, dx)
+para.set_has_wall_model_monitor(True)
+grid_builder.set_slip_boundary_condition(gpu.SideType.PZ, 0, 0, -1)
+
+if read_precursor:
+ grid_builder.set_pressure_boundary_condition(gpu.SideType.PX, 0)
+bc_factory.set_stress_boundary_condition(gpu.StressBC.StressPressureBounceBack)
+bc_factory.set_slip_boundary_condition(gpu.SlipBC.SlipBounceBack)
+bc_factory.set_pressure_boundary_condition(gpu.PressureBC.OutflowNonReflective)
+if read_precursor:
+ bc_factory.set_precursor_boundary_condition(gpu.PrecursorBC.DistributionsPrecursor if use_distributions else gpu.PrecursorBC.VelocityPrecursor)
+para.set_outflow_pressure_correction_factor(0.0);
#%%
-cuda_memory_manager = gpu.CudaMemoryManager(para)
-grid_generator = gpu.GridProvider.make_grid_generator(grid_builder, para, cuda_memory_manager, comm)
+# don't use python init functions, they are very slow! Just kept as an example.
+# Define lambda in bindings and set it here.
+# def init_func(coord_x, coord_y, coord_z):
+# return [
+# 0.0,
+# (u_star/0.4 * np.log(np.maximum(coord_z,z0)/z0) + 2.0*np.sin(np.pi*16*coord_x/length[0])*np.sin(np.pi*8*coord_z/boundary_layer_height)/(np.square(coord_z/boundary_layer_height)+1)) * dt / dx,
+# 2.0*np.sin(np.pi*16.*coord_x/length[0])*np.sin(np.pi*8.*coord_z/boundary_layer_height)/(np.square(coord_z/boundary_layer_height)+1.) * dt / dx,
+# 8.0*u_star/0.4*(np.sin(np.pi*8.0*coord_y/boundary_layer_height)*np.sin(np.pi*8.0*coord_z/boundary_layer_height)+np.sin(np.pi*8.0*coord_x/length[0]))/(np.square(length[2]/2.0-coord_z)+1.) * dt / dx]
+# para.set_initial_condition(init_func)
+para.set_initial_condition_perturbed_log_law(u_star, z0, length[0], length[2], boundary_layer_height, velocity_ratio)
+
#%%
-turb_pos = np.array([3,3,3])*reference_diameter
-epsilon = 5
+turb_pos = np.array([3,3,3])*turbine_diameter
+epsilon = 1.5*dx
density = 1.225
level = 0
n_blades = 3
n_blade_nodes = 32
-alm = gpu.ActuatorLine(n_blades, density, n_blade_nodes, epsilon, *turb_pos, reference_diameter, level, dt, dx)
+omega = 1
+blade_radii = np.arange(n_blade_nodes, dtype=np.float32)/(0.5*turbine_diameter)
+alm = gpu.ActuatorFarm(n_blades, density, n_blade_nodes, epsilon, level, dt, dx, True)
+alm.add_turbine(turb_pos[0],turb_pos[1],turb_pos[2], turbine_diameter, omega, 0, 0, blade_radii)
para.add_actuator(alm)
#%%
-point_probe = gpu.probes.PointProbe("pointProbe", str(output_path), 100, 1, 500, 100)
-point_probe.add_probe_points_from_list(np.array([1,2,5])*reference_diameter, np.array([3,3,3])*reference_diameter, np.array([3,3,3])*reference_diameter)
-point_probe.add_statistic(gpu.probes.Statistic.Means)
-
-para.add_probe(point_probe)
-
-plane_probe = gpu.probes.PlaneProbe("planeProbe", str(output_path), 100, 1, 500, 100)
-plane_probe.set_probe_plane(5*reference_diameter, 0, 0, dx, length[1], length[2])
-para.add_probe(plane_probe)
+planar_average_probe = gpu.probes.PlanarAverageProbe("horizontalPlanes", para.get_output_path(), 0, int(t_start_tmp_averaging/dt), int(t_averaging/dt) , int(t_start_out_probe/dt), int(t_out_probe/dt), 'z')
+planar_average_probe.add_all_available_statistics()
+planar_average_probe.set_file_name_to_n_out()
+para.add_probe(planar_average_probe)
#%%
-sim = gpu.Simulation(para, cuda_memory_manager, comm, grid_generator)
+wall_model_probe = gpu.probes.WallModelProbe("wallModelProbe", para.get_output_path(), 0, int(t_start_tmp_averaging/dt), int(t_averaging/dt/4), int(t_start_out_probe/dt), int(t_out_probe/dt))
+wall_model_probe.add_all_available_statistics()
+wall_model_probe.set_file_name_to_n_out()
+wall_model_probe.set_force_output_to_stress(True)
+if para.get_is_body_force():
+ wall_model_probe.set_evaluate_pressure_gradient(True)
+para.add_probe(wall_model_probe)
+
+plane_locs = [100,]
+if read_precursor: plane_locs.extend([1000, 1500, 2000, 2500, 0])
+
+for n_probe, probe_pos in enumerate(plane_locs):
+ plane_probe = gpu.probes.PlaneProbe(f"planeProbe_{n_probe+1}", para.get_output_path(), int(t_start_averaging/dt), 10, int(t_start_out_probe/dt), int(t_out_probe/dt))
+ plane_probe.set_probe_plane(probe_pos, 0, 0, dx, length[1], length[2])
+ plane_probe.add_all_available_statistics()
+ para.add_probe(plane_probe)
+#%%
+cuda_memory_manager = gpu.CudaMemoryManager(para)
+grid_generator = gpu.GridProvider.make_grid_generator(grid_builder, para, cuda_memory_manager, communicator)
+#%%
+#%%
+sim = gpu.Simulation(para, cuda_memory_manager, communicator, grid_generator, bc_factory, tm_factory, grid_scaling_factory)
#%%
sim.run()
-MPI.Finalize()
\ No newline at end of file
+MPI.Finalize()
+
diff --git a/Python/actuator_line/config.txt b/Python/actuator_line/config.txt
deleted file mode 100644
index e4c778c4cc048f54c0a32310e6bf4a7343a263fa..0000000000000000000000000000000000000000
--- a/Python/actuator_line/config.txt
+++ /dev/null
@@ -1,2 +0,0 @@
-Path = .
-GridPath = .
diff --git a/Python/actuator_line/configActuatorLine.txt b/Python/actuator_line/configActuatorLine.txt
new file mode 100644
index 0000000000000000000000000000000000000000..c45d170f039274ab355f3fe1dc044536f1f29e6f
--- /dev/null
+++ b/Python/actuator_line/configActuatorLine.txt
@@ -0,0 +1,39 @@
+##################################################
+#informations for Writing
+##################################################
+Path = .
+##################################################
+#informations for reading
+##################################################
+GridPath = .
+##################################################
+Devices = 0
+##################################################
+tStartOut = 0
+tOut = 100000
+tEnd = 300000
+##################################################
+tStartAveraging = 0
+tStartTmpAveraging = 100000
+tAveraging = 200
+tStartOutProbe = 0
+tOutProbe = 1000
+##################################################
+Ma = 0.1
+nz = 96
+
+bodyForce = true
+SGSconstant = 0.333
+TurbulenceModel = QR
+
+QuadricLimiterP = 100000.0
+QuadricLimiterM = 100000.0
+QuadricLimiterD = 100000.0
+
+##################################################
+readPrecursor = false
+nTimestepsReadPrecursor = 1
+precursorFile = precursor/Precursor
+
+##################################################
+turbineDiameter = 126.0
diff --git a/Python/boundary_layer/boundary_layer.py b/Python/boundary_layer/boundary_layer.py
index 1c01f50946b49bc0ddab7e50065a24aab4ae869f..6f6c64bc072d3afbb8aa5febbec209c26af2deee 100644
--- a/Python/boundary_layer/boundary_layer.py
+++ b/Python/boundary_layer/boundary_layer.py
@@ -1,37 +1,48 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file boundary_layer.py
+! \ingroup boundary_layer
+! \author Henry Korb, Henrik Asmuth
+=======================================================================================
+"""
#%%
import numpy as np
from pathlib import Path
from mpi4py import MPI
from pyfluids import basics, gpu, logger
#%%
-reference_height = 1000 # boundary layer height in m
-
-length = np.array([6,4,1])*reference_height
-viscosity = 1.56e-5
-mach = 0.1
-nodes_per_height = 32
-
-z_0 = 0.1
-u_star = 0.4
-kappa = 0.4
-
-velocity = 0.5*u_star/kappa*np.log(length[2]/z_0+1)
-flow_through_time = length[0]/velocity
-use_AMD = True
-
-
-sim_name = "BoundaryLayer"
-config_file = Path(__file__).parent/Path("config.txt")
+sim_name = "ABL"
+config_file = Path(__file__).parent/"configBoundaryLayer.txt"
output_path = Path(__file__).parent/Path("output")
output_path.mkdir(exist_ok=True)
-t_out = 1000.
-t_end = 5000.
-t_start_averaging = 0
-t_start_tmp_averaging = 100_000
-t_averaging = 200
-t_start_out_probe = 0
-t_out_probe = 1000
#%%
logger.Logger.initialize_logger()
@@ -39,95 +50,161 @@ basics.logger.Logger.add_stdout()
basics.logger.Logger.set_debug_level(basics.logger.Level.INFO_LOW)
basics.logger.Logger.time_stamp(basics.logger.TimeStamp.ENABLE)
basics.logger.Logger.enable_printed_rank_numbers(True)
-# %%
-comm = gpu.Communicator.get_instance()
#%%
grid_factory = gpu.grid_generator.GridFactory.make()
grid_builder = gpu.grid_generator.MultipleGridBuilder.make_shared(grid_factory)
+communicator = gpu.Communicator.get_instance()
+
+config = basics.ConfigurationFile()
+config.load(str(config_file))
+
+para = gpu.Parameter(communicator.get_number_of_process(), communicator.get_pid(), config)
+bc_factory = gpu.BoundaryConditionFactory()
#%%
-dx = reference_height/nodes_per_height
-dt = dx * mach / (np.sqrt(3) * velocity)
-velocity_lb = velocity * dt / dx # LB units
-viscosity_lb = viscosity * dt / (dx * dx) # LB units
+boundary_layer_height = config.get_float_value("boundaryLayerHeight", 1000)
+z0 = config.get_float_value("z0", 0.1)
+u_star = config.get_float_value("u_star", 0.4)
-pressure_gradient = u_star**2 / reference_height
-pressure_gradient_lb = pressure_gradient * dt**2 / dx
+kappa = config.get_float_value("vonKarmanConstant", 0.4) # von Karman constant
-logger.vf_log_info(f"velocity = {velocity_lb:1.6} dx/dt")
-logger.vf_log_info(f"dt = {dt:1.6}")
-logger.vf_log_info(f"dx = {dx:1.6}")
-logger.vf_log_info(f"u* = {u_star:1.6}")
-logger.vf_log_info(f"dpdx = {pressure_gradient:1.6}")
-logger.vf_log_info(f"dpdx = {pressure_gradient_lb:1.6} dx/dt^2")
-logger.vf_log_info(f"viscosity = {viscosity_lb:1.6} dx^2/dt")
+viscosity = config.get_float_value("viscosity", 1.56e-5)
+velocity = 0.5*u_star/kappa*np.log(boundary_layer_height/z0+1) #0.5 times max mean velocity at the top in m/s
-#%%
-config = basics.ConfigurationFile()
-config.load(str(config_file))
-#%%
-para = gpu.Parameter(config, comm.get_number_of_process(), comm.get_pid())
+mach = config.get_float_value("Ma", 0.1)
+nodes_per_height = config.get_uint_value("nz", 64)
+
+
+
+write_precursor = config.get_bool_value("_p", False)
+read_precursor = config.get_bool_value("readPrecursor", False)
+
+if write_precursor:
+ nTWritePrecursor = config.get_int_value("nTimestepsWritePrecursor")
+ t_start_precursor = config.get_float_value("tStartPrecursor")
+ pos_x_precursor = config.get_float_value("posXPrecursor")
+if read_precursor:
+ nTReadPrecursor = config.get_int_value("nTimestepsReadPrecursor")
+if write_precursor or read_precursor:
+ use_distributions = config.get_bool_value("useDistributions", False)
+ precursor_directory = config.get_string_value("precursorDirectory")
+
+# all in s
+t_start_out = config.get_float_value("tStartOut")
+t_out = config.get_float_value("tOut")
+t_end = config.get_float_value("tEnd") # total time of simulation
+
+t_start_averaging = config.get_float_value("tStartAveraging")
+t_start_tmp_averaging = config.get_float_value("tStartTmpAveraging")
+t_averaging = config.get_float_value("tAveraging")
+t_start_out_probe = config.get_float_value("tStartOutProbe")
+t_out_probe = config.get_float_value("tOutProbe")
+
+#%%
+length = np.array([6,4,1])*boundary_layer_height
+dx = boundary_layer_height/nodes_per_height
+dt = dx * mach / (np.sqrt(3) * velocity)
+velocity_LB = velocity * dt / dx # LB units
+viscosity_LB = viscosity * dt / (dx * dx) # LB units
+pressure_gradient = u_star * u_star / boundary_layer_height
+pressure_gradient_LB = pressure_gradient * (dt*dt)/dx
+
+logger.vf_log_info(f"velocity [dx/dt] = {velocity_LB}")
+logger.vf_log_info(f"dt = {dt}")
+logger.vf_log_info(f"dx = {dx}")
+logger.vf_log_info(f"viscosity [10^8 dx^2/dt] = {viscosity_LB*1e8}")
+logger.vf_log_info(f"u* /(dx/dt) = {u_star*dt/dx}")
+logger.vf_log_info(f"dpdx = {pressure_gradient}")
+logger.vf_log_info(f"dpdx /(dx/dt^2) = {pressure_gradient_LB}")
+
+#%%
#%%
-para.set_devices([0])
para.set_output_prefix(sim_name)
-para.set_output_path(str(output_path))
-para.set_f_name(para.get_output_path() + "/" + para.get_output_prefix())
para.set_print_files(True)
-para.set_max_level(1)
-#%%
-para.set_velocity(velocity_lb)
-para.set_viscosity(viscosity_lb)
+
+para.set_forcing(pressure_gradient_LB, 0, 0)
+para.set_velocity_LB(velocity_LB)
+para.set_viscosity_LB(viscosity_LB)
para.set_velocity_ratio(dx/dt)
para.set_viscosity_ratio(dx*dx/dt)
-para.set_use_AMD(use_AMD)
+para.set_density_ratio(1.0)
+
+para.set_main_kernel("CumulantK17")
-para.set_main_kernel("TurbulentViscosityCumulantK17CompChim" if para.get_use_AMD() else "CummulantK17CompChim")
+para.set_timestep_start_out(int(t_start_out/dt))
+para.set_timestep_out(int(t_out/dt))
+para.set_timestep_end(int(t_end/dt))
+para.set_is_body_force(config.get_bool_value("bodyForce"))
+#%%
+tm_factory = gpu.TurbulenceModelFactory(para)
+tm_factory.read_config_file(config)
+#%%
+grid_builder.add_coarse_grid(0.0, 0.0, 0.0, *length, dx)
+grid_builder.set_periodic_boundary_condition(not read_precursor, True, False)
+grid_builder.build_grids(basics.LbmOrGks.LBM, False)
-para.set_SGS_constant(0.083)
+sampling_offset = 2
+if read_precursor:
+ precursor = gpu.create_file_collection(precursor_directory + "/precursor", gpu.FileType.VTK)
+ grid_builder.set_precursor_boundary_condition(gpu.SideType.MX, precursor, nTReadPrecursor, 0, 0, 0)
+grid_builder.set_stress_boundary_condition(gpu.SideType.MZ, 0, 0, 1, sampling_offset, z0/dx)
+para.set_has_wall_model_monitor(True)
+grid_builder.set_slip_boundary_condition(gpu.SideType.PZ, 0, 0, -1)
+
+if read_precursor:
+ grid_builder.set_pressure_boundary_condition(gpu.SideType.PX, 0)
+bc_factory.set_stress_boundary_condition(gpu.StressBC.StressPressureBounceBack)
+bc_factory.set_slip_boundary_condition(gpu.SlipBC.SlipBounceBack)
+bc_factory.set_pressure_boundary_condition(gpu.PressureBC.OutflowNonReflective)
+bc_factory.set_precursor_boundary_condition(gpu.PrecursorBC.DistributionsPrecursor if use_distributions else gpu.PrecursorBC.VelocityPrecursor)
+para.set_outflow_pressure_correction_factor(0.0);
+#%%
def init_func(coord_x, coord_y, coord_z):
return [
0.0,
- (u_star/kappa*np.log(max(coord_z/z_0,0)+1) + 2*np.sin(np.pi*16*coord_x/length[0])*np.sin(np.pi*8*coord_z/length[2]))/((coord_z/reference_height)**2+0.1)*dt/dx,
- 2*np.sin(np.pi*16*coord_x/length[0])*np.sin(np.pi*8*coord_z/length[2])/((coord_z/reference_height)**2+0.1)*dt/dx,
- 8*u_star/kappa*(np.sin(np.pi*8*coord_y/reference_height)*np.sin(np.pi*8*coord_z/reference_height)+np.sin(np.pi*8*coord_x/length[0]))/((length[2]/2-coord_z)**2+0.1)*dt/dx
- ]
-
+ (u_star/0.4 * np.log(np.maximum(coord_z,z0)/z0) + 2.0*np.sin(np.pi*16*coord_x/length[0])*np.sin(np.pi*8*coord_z/boundary_layer_height)/(np.square(coord_z/boundary_layer_height)+1)) * dt / dx,
+ 2.0*np.sin(np.pi*16.*coord_x/length[0])*np.sin(np.pi*8.*coord_z/boundary_layer_height)/(np.square(coord_z/boundary_layer_height)+1.) * dt / dx,
+ 8.0*u_star/0.4*(np.sin(np.pi*8.0*coord_y/boundary_layer_height)*np.sin(np.pi*8.0*coord_z/boundary_layer_height)+np.sin(np.pi*8.0*coord_x/length[0]))/(np.square(length[2]/2.0-coord_z)+1.) * dt / dx]
para.set_initial_condition(init_func)
-para.set_t_out(int(t_out/dt))
-para.set_t_end(int(t_end/dt))
-para.set_is_body_force(True)
-para.set_has_wall_model_monitor(True)
-
-grid_builder.add_coarse_grid(0.0, 0.0, 0.0, *length, dx)
-grid_builder.set_periodic_boundary_condition(True, True, False)
-grid_builder.build_grids(basics.LbmOrGks.LBM, False)
#%%
-sampling_offset = 2
-grid_builder.set_stress_boundary_condition(gpu.SideType.MZ, 0.0, 0.0, 1.0, sampling_offset, z_0/dx)
-grid_builder.set_slip_boundary_condition(gpu.SideType.PZ, 0.0, 0.0, 0.0)
+planar_average_probe = gpu.probes.PlanarAverageProbe("horizontalPlanes", para.get_output_path(), 0, int(t_start_tmp_averaging/dt), int(t_averaging/dt) , int(t_start_out_probe/dt), int(t_out_probe/dt), 'z')
+planar_average_probe.add_all_available_statistics()
+planar_average_probe.set_file_name_to_n_out()
+para.add_probe(planar_average_probe)
+#%%
+wall_model_probe = gpu.probes.WallModelProbe("wallModelProbe", para.get_output_path(), 0, int(t_start_tmp_averaging/dt), int(t_averaging/dt/4), int(t_start_out_probe/dt), int(t_out_probe/dt))
+wall_model_probe.add_all_available_statistics()
+wall_model_probe.set_file_name_to_n_out()
+wall_model_probe.set_force_output_to_stress(True)
+if para.get_is_body_force():
+ wall_model_probe.set_evaluate_pressure_gradient(True)
+para.add_probe(wall_model_probe)
+
+plane_locs = [100,]
+if read_precursor: plane_locs.extend([1000, 1500, 2000, 2500, 0])
+
+for n_probe, probe_pos in enumerate(plane_locs):
+ plane_probe = gpu.probes.PlaneProbe(f"planeProbe_{n_probe+1}", para.get_output_path(), int(t_start_averaging/dt), 10, int(t_start_out_probe/dt), int(t_out_probe/dt))
+ plane_probe.set_probe_plane(probe_pos, 0, 0, dx, length[1], length[2])
+ plane_probe.add_all_available_statistics()
+ para.add_probe(plane_probe)
+
+if write_precursor:
+ precursor_writer = gpu.PrecursorWriter("precursor", para.get_output_path() + precursor_directory, pos_x_precursor, 0,length[1], 0, length[2], t_start_precursor/dt, nTWritePrecursor, gpu.OutputVariable.Distributions if use_distributions else gpu.OutputVariable.Velocities)
+ para.add_probe(precursor_writer)
#%%
cuda_memory_manager = gpu.CudaMemoryManager(para)
-grid_generator = gpu.GridProvider.make_grid_generator(grid_builder, para, cuda_memory_manager, comm)
-
+grid_generator = gpu.GridProvider.make_grid_generator(grid_builder, para, cuda_memory_manager, communicator)
#%%
-wall_probe = gpu.probes.WallModelProbe("wallModelProbe", str(output_path), int(t_start_averaging/dt), int(t_start_tmp_averaging/dt), int(t_averaging/dt/4), int(t_start_out_probe/dt), int(t_out_probe/dt))
-wall_probe.add_all_available_statistics()
-wall_probe.set_file_name_to_n_out()
-wall_probe.set_force_output_to_stress(True)
-if para.get_is_body_force():
- wall_probe.set_evaluate_pressure_gradient(True)
-planar_probe = gpu.probes.PlanarAverageProbe("planarAverageProbe", str(output_path), int(t_start_averaging/dt), int(t_start_tmp_averaging/dt), int(t_averaging/dt), int(t_start_out_probe/dt), int(t_out_probe/dt), "z")
-para.add_probe(wall_probe)
-
#%%
-sim = gpu.Simulation(para, cuda_memory_manager, comm, grid_generator)
+sim = gpu.Simulation(para, cuda_memory_manager, communicator, grid_generator, bc_factory, tm_factory)
#%%
sim.run()
MPI.Finalize()
\ No newline at end of file
diff --git a/Python/boundary_layer/config.txt b/Python/boundary_layer/config.txt
deleted file mode 100644
index e4c778c4cc048f54c0a32310e6bf4a7343a263fa..0000000000000000000000000000000000000000
--- a/Python/boundary_layer/config.txt
+++ /dev/null
@@ -1,2 +0,0 @@
-Path = .
-GridPath = .
diff --git a/Python/boundary_layer/configBoundaryLayer.txt b/Python/boundary_layer/configBoundaryLayer.txt
new file mode 100644
index 0000000000000000000000000000000000000000..83e7861a5fb85ea800d187699f1c6c1409422f0a
--- /dev/null
+++ b/Python/boundary_layer/configBoundaryLayer.txt
@@ -0,0 +1,42 @@
+##################################################
+#informations for Writing
+##################################################
+Path = .
+##################################################
+#informations for reading
+##################################################
+GridPath = .
+##################################################
+Devices = 0
+##################################################
+tStartOut = 0
+tOut = 100000
+tEnd = 300000
+##################################################
+tStartAveraging = 0
+tStartTmpAveraging = 100000
+tAveraging = 200
+tStartOutProbe = 0
+tOutProbe = 1000
+##################################################
+Ma = 0.1
+nz = 96
+
+bodyForce = true
+UseAMD = true
+SGSconstant = 0.2
+QuadricLimiterP = 100000.0
+QuadricLimiterM = 100000.0
+QuadricLimiterD = 100000.0
+
+##################################################
+readPrecursor = false
+nTimestepsReadPrecursor = 10
+precursorFile = precursor/Precursor
+
+##################################################
+writePrecursor = false
+nTimestepsWritePrecursor = 10
+
+tStartPrecursor = 100
+posXPrecursor = 3000
\ No newline at end of file
diff --git a/Python/cubeflow/simulation.py b/Python/cubeflow/simulation.py
index 9e77e8d747c072188d8d81150afa8e2ccb76a792..deb0411963aec65522af45cc48d7367f103232c6 100644
--- a/Python/cubeflow/simulation.py
+++ b/Python/cubeflow/simulation.py
@@ -1,13 +1,42 @@
-from pyfluids.cpu import Simulation
-from pyfluids.cpu.boundaryconditions import NoSlipBoundaryCondition, VelocityBoundaryCondition, DensityBoundaryCondition
-from pyfluids.cpu.geometry import GbCuboid3D
-from pyfluids.cpu.kernel import LBMKernel, KernelType
-from pyfluids.cpu.parameters import PhysicalParameters, RuntimeParameters, GridParameters
-from pyfluids.cpu.writer import Writer, OutputFormat
-from pymuparser import Parser
-
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file simulation.py
+! \ingroup cubeflow
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
import os
+from pyfluids import cpu
+from pymuparser import Parser
+
def get_max_length(number_of_nodes_per_direction, delta_x):
return (number_of_nodes_per_direction[0] * delta_x,
@@ -15,10 +44,10 @@ def get_max_length(number_of_nodes_per_direction, delta_x):
number_of_nodes_per_direction[2] * delta_x)
-physical_params = PhysicalParameters()
+physical_params = cpu.parameters.PhysicalParameters()
physical_params.lattice_viscosity = 0.005
-grid_params = GridParameters()
+grid_params = cpu.parameters.GridParameters()
grid_params.number_of_nodes_per_direction = [200, 120, 120]
grid_params.blocks_per_direction = [2, 2, 2]
grid_params.node_distance = 0.125
@@ -26,7 +55,7 @@ grid_params.periodic_boundary_in_x1 = False
grid_params.periodic_boundary_in_x2 = True
grid_params.periodic_boundary_in_x3 = True
-runtime_params = RuntimeParameters()
+runtime_params = cpu.parameters.RuntimeParameters()
runtime_params.timestep_log_interval = 1000
runtime_params.number_of_timesteps = 50000
runtime_params.number_of_threads = int(os.environ.get("OMP_NUM_THREADS", 4))
@@ -39,46 +68,46 @@ def run_simulation(physical_parameters=physical_params, grid_parameters=grid_par
min_x, min_y, min_z = 0, 0, 0
max_x, max_y, max_z = get_max_length(grid_parameters.number_of_nodes_per_direction, grid_parameters.node_distance)
- bottom_wall = GbCuboid3D(min_x - wall_thickness, min_y - wall_thickness, min_z, max_x + wall_thickness,
+ bottom_wall = cpu.geometry.GbCuboid3D(min_x - wall_thickness, min_y - wall_thickness, min_z, max_x + wall_thickness,
max_y + wall_thickness, min_z - wall_thickness)
- top_wall = GbCuboid3D(min_x - wall_thickness, min_y - wall_thickness, max_z, max_x + wall_thickness,
+ top_wall = cpu.geometry.GbCuboid3D(min_x - wall_thickness, min_y - wall_thickness, max_z, max_x + wall_thickness,
max_y + wall_thickness,
max_z + wall_thickness)
- left_wall = GbCuboid3D(min_x - wall_thickness, min_y, min_z - wall_thickness, max_x + wall_thickness,
+ left_wall = cpu.geometry.GbCuboid3D(min_x - wall_thickness, min_y, min_z - wall_thickness, max_x + wall_thickness,
min_y - wall_thickness,
max_z + wall_thickness)
- right_wall = GbCuboid3D(min_x - wall_thickness, max_y, min_z - wall_thickness, max_x + wall_thickness,
+ right_wall = cpu.geometry.GbCuboid3D(min_x - wall_thickness, max_y, min_z - wall_thickness, max_x + wall_thickness,
max_y + wall_thickness, max_z + wall_thickness)
- obstacle = GbCuboid3D(7, 7, 7, 8, 8, 8)
+ obstacle = cpu.geometry.GbCuboid3D(7, 7, 7, 8, 8, 8)
- velocity_boundary = GbCuboid3D(min_x - wall_thickness, min_y - wall_thickness, min_z - wall_thickness, min_x,
+ velocity_boundary = cpu.geometry.GbCuboid3D(min_x - wall_thickness, min_y - wall_thickness, min_z - wall_thickness, min_x,
max_y + wall_thickness, max_z + wall_thickness)
- outflow_boundary = GbCuboid3D(max_x, min_y - wall_thickness, min_z - wall_thickness, max_x + wall_thickness,
+ outflow_boundary = cpu.geometry.GbCuboid3D(max_x, min_y - wall_thickness, min_z - wall_thickness, max_x + wall_thickness,
max_y + wall_thickness, max_z + wall_thickness)
- no_slip_bc = NoSlipBoundaryCondition()
+ no_slip_bc = cpu.boundaryconditions.NoSlipBoundaryCondition()
- outflow_bc = DensityBoundaryCondition()
+ outflow_bc = cpu.boundaryconditions.DensityBoundaryCondition()
velocity_function = Parser()
velocity_function.define_constant("u", 0.07)
velocity_function.expression = "u"
- velocity_bc = VelocityBoundaryCondition(True, False, False, velocity_function, 0, -10)
+ velocity_bc = cpu.boundaryconditions.VelocityBoundaryCondition(True, False, False, velocity_function, 0, -10)
- kernel = LBMKernel(KernelType.CompressibleCumulantFourthOrderViscosity)
+ kernel = cpu.kernel.LBMKernel(cpu.kernel.KernelType.CompressibleCumulantFourthOrderViscosity)
# kernel.use_forcing = True
# kernel.forcing_in_x1 = 3e-6
- writer = Writer()
+ writer = cpu.writer.Writer()
writer.output_path = "./output"
- writer.output_format = OutputFormat.BINARY
+ writer.output_format = cpu.writer.OutputFormat.BINARY
- simulation = Simulation()
+ simulation = cpu.Simulation()
simulation.set_writer(writer)
simulation.set_physical_parameters(physical_parameters)
diff --git a/Python/liddrivencavity/simulation.py b/Python/liddrivencavity/simulation.py
index 155fad2f6f8aade0368c8a7006b88f7985f8822c..3c247b87a102e3c5a720f20748acc9f9f50bb178 100644
--- a/Python/liddrivencavity/simulation.py
+++ b/Python/liddrivencavity/simulation.py
@@ -1,32 +1,61 @@
-from pyfluids.cpu import Simulation
-from pyfluids.cpu.boundaryconditions import NoSlipBoundaryCondition, VelocityBoundaryCondition
-from pyfluids.cpu.geometry import GbCuboid3D
-from pyfluids.cpu.kernel import LBMKernel, KernelType
-from pyfluids.cpu.parameters import GridParameters, PhysicalParameters, RuntimeParameters
-from pyfluids.cpu.writer import Writer, OutputFormat
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file simulation.py
+! \ingroup liddrivencavity
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
+from pyfluids import cpu
from pymuparser import Parser
-runtime_params = RuntimeParameters()
+runtime_params = cpu.parameters.RuntimeParameters()
runtime_params.number_of_threads = 4
runtime_params.number_of_timesteps = 10000
runtime_params.timestep_log_interval = 1000
-physical_params = PhysicalParameters()
+physical_params = cpu.parameters.PhysicalParameters()
physical_params.lattice_viscosity = 0.005
-grid_params = GridParameters()
+grid_params = cpu.parameters.GridParameters()
grid_params.number_of_nodes_per_direction = [64, 64, 64]
grid_params.blocks_per_direction = [2, 2, 2]
grid_params.node_distance = 1 / 10
def run_simulation(physical_params=physical_params, grid_params=grid_params, runtime_params=runtime_params):
- simulation = Simulation()
- kernel = LBMKernel(KernelType.CompressibleCumulantFourthOrderViscosity)
+ simulation = cpu.Simulation()
+ kernel = cpu.kernel.LBMKernel(cpu.kernel.KernelType.CompressibleCumulantFourthOrderViscosity)
- writer = Writer()
+ writer = cpu.writer.Writer()
writer.output_path = "./output"
- writer.output_format = OutputFormat.BINARY
+ writer.output_format = cpu.writer.OutputFormat.BINARY
simulation.set_grid_parameters(grid_params)
simulation.set_physical_parameters(physical_params)
@@ -34,12 +63,12 @@ def run_simulation(physical_params=physical_params, grid_params=grid_params, run
simulation.set_kernel_config(kernel)
simulation.set_writer(writer)
- no_slip_bc_adapter = NoSlipBoundaryCondition()
+ no_slip_bc_adapter = cpu.boundaryconditions.NoSlipBoundaryCondition()
fct = Parser()
fct.expression = "u"
fct.define_constant("u", 0.005)
- velocity_bc_adapter = VelocityBoundaryCondition(True, True, False, fct, 0, -10.0)
+ velocity_bc_adapter = cpu.boundaryconditions.VelocityBoundaryCondition(True, True, False, fct, 0, -10.0)
g_min_x1, g_min_x2, g_min_x3 = 0, 0, 0
g_max_x1 = grid_params.number_of_nodes_per_direction[0] * grid_params.node_distance
@@ -48,12 +77,12 @@ def run_simulation(physical_params=physical_params, grid_params=grid_params, run
dx = grid_params.node_distance
- wall_x_min = GbCuboid3D(g_min_x1 - dx, g_min_x2 - dx, g_min_x3 - dx, g_min_x1, g_max_x2 + dx, g_max_x3)
- wall_x_max = GbCuboid3D(g_max_x1, g_min_x2 - dx, g_min_x3 - dx, g_max_x1 + dx, g_max_x2 + dx, g_max_x3)
- wall_y_min = GbCuboid3D(g_min_x1 - dx, g_min_x2 - dx, g_min_x3 - dx, g_max_x1 + dx, g_min_x2, g_max_x3)
- wall_y_max = GbCuboid3D(g_min_x1 - dx, g_max_x2, g_min_x3 - dx, g_max_x1 + dx, g_max_x2 + dx, g_max_x3)
- wall_z_min = GbCuboid3D(g_min_x1 - dx, g_min_x2 - dx, g_min_x3 - dx, g_max_x1 + dx, g_max_x2 + dx, g_min_x3)
- wall_z_max = GbCuboid3D(g_min_x1 - dx, g_min_x2 - dx, g_max_x3, g_max_x1 + dx, g_max_x2 + dx, g_max_x3 + dx)
+ wall_x_min = cpu.geometry.GbCuboid3D(g_min_x1 - dx, g_min_x2 - dx, g_min_x3 - dx, g_min_x1, g_max_x2 + dx, g_max_x3)
+ wall_x_max = cpu.geometry.GbCuboid3D(g_max_x1, g_min_x2 - dx, g_min_x3 - dx, g_max_x1 + dx, g_max_x2 + dx, g_max_x3)
+ wall_y_min = cpu.geometry.GbCuboid3D(g_min_x1 - dx, g_min_x2 - dx, g_min_x3 - dx, g_max_x1 + dx, g_min_x2, g_max_x3)
+ wall_y_max = cpu.geometry.GbCuboid3D(g_min_x1 - dx, g_max_x2, g_min_x3 - dx, g_max_x1 + dx, g_max_x2 + dx, g_max_x3)
+ wall_z_min = cpu.geometry.GbCuboid3D(g_min_x1 - dx, g_min_x2 - dx, g_min_x3 - dx, g_max_x1 + dx, g_max_x2 + dx, g_min_x3)
+ wall_z_max = cpu.geometry.GbCuboid3D(g_min_x1 - dx, g_min_x2 - dx, g_max_x3, g_max_x1 + dx, g_max_x2 + dx, g_max_x3 + dx)
simulation.add_object(wall_x_min, no_slip_bc_adapter, 1, "/geo/wallXmin")
simulation.add_object(wall_x_max, no_slip_bc_adapter, 1, "/geo/wallXmax")
diff --git a/Python/poiseuille/poiseuille_hpc.py b/Python/poiseuille/poiseuille_hpc.py
index f5f5a1387c9fe234abae0c6f979cc7d5b283d1a4..b108f34445a71a686c4e22f685e26e10204113b3 100644
--- a/Python/poiseuille/poiseuille_hpc.py
+++ b/Python/poiseuille/poiseuille_hpc.py
@@ -1,15 +1,49 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file poiseuille_hpc.py
+! \ingroup poiseuille
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
from poiseuille.simulation import run_simulation
-from pyfluids.cpu.parameters import *
+from pyfluids import cpu
-grid_parameters = GridParameters()
+grid_parameters = cpu.prameters.GridParameters()
grid_parameters.number_of_nodes_per_direction = [64, 64, 512]
grid_parameters.node_distance = 1
grid_parameters.blocks_per_direction = [1, 2, 2]
-physical_parameters = PhysicalParameters()
+physical_parameters = cpu.prameters.PhysicalParameters()
physical_parameters.lattice_viscosity = 0.0005
-runtime_parameters = RuntimeParameters()
+runtime_parameters = cpu.prameters.RuntimeParameters()
runtime_parameters.number_of_threads = 4
runtime_parameters.number_of_timesteps = 1000
runtime_parameters.timestep_log_interval = 100
diff --git a/Python/poiseuille/simulation.py b/Python/poiseuille/simulation.py
index d107801fa84cfe16d1d7e91d31dc3ff4b8671f02..a6f12e59fbd0a0ccad9a4db9ccde69b828cf90bf 100644
--- a/Python/poiseuille/simulation.py
+++ b/Python/poiseuille/simulation.py
@@ -1,35 +1,65 @@
-from pyfluids.cpu import Simulation
-from pyfluids.cpu.boundaryconditions import NoSlipBoundaryCondition
-from pyfluids.cpu.geometry import GbCuboid3D, State
-from pyfluids.cpu.kernel import LBMKernel, KernelType
-from pyfluids.cpu.parameters import RuntimeParameters, GridParameters, PhysicalParameters
-from pyfluids.cpu.writer import Writer, OutputFormat
-
-default_grid_params = GridParameters()
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file simulation.py
+! \ingroup poiseuille
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
+from pyfluids import cpu
+
+
+default_grid_params = cpu.parameters.GridParameters()
default_grid_params.node_distance = 10 / 32
default_grid_params.number_of_nodes_per_direction = [8, 8, 32]
default_grid_params.blocks_per_direction = [1, 1, 4]
default_grid_params.periodic_boundary_in_x1 = True
default_grid_params.periodic_boundary_in_x2 = True
-default_physical_params = PhysicalParameters()
+default_physical_params = cpu.parameters.PhysicalParameters()
default_physical_params.lattice_viscosity = 0.005
-default_runtime_params = RuntimeParameters()
+default_runtime_params = cpu.parameters.RuntimeParameters()
default_runtime_params.number_of_threads = 4
default_runtime_params.number_of_timesteps = 10000
default_runtime_params.timestep_log_interval = 1000
-default_kernel = LBMKernel(KernelType.CompressibleCumulantFourthOrderViscosity)
+default_kernel = cpu.kernel.LBMKernel(cpu.kernel.KernelType.CompressibleCumulantFourthOrderViscosity)
default_kernel.use_forcing = True
default_kernel.forcing_in_x1 = 1e-8
-default_writer = Writer()
+default_writer = cpu.writer.Writer()
default_writer.output_path = "./output"
-default_writer.output_format = OutputFormat.BINARY
+default_writer.output_format = cpu.writer.OutputFormat.BINARY
-default_kernel = LBMKernel(KernelType.CompressibleCumulantFourthOrderViscosity)
+default_kernel = cpu.kernel.LBMKernel(cpu.kernel.KernelType.CompressibleCumulantFourthOrderViscosity)
default_kernel.use_forcing = True
default_kernel.forcing_in_x1 = 1e-8
@@ -39,7 +69,7 @@ def run_simulation(physical_params=default_physical_params,
runtime_params=default_runtime_params,
kernel=default_kernel,
writer=default_writer):
- simulation = Simulation()
+ simulation = cpu.Simulation()
simulation.set_kernel_config(kernel)
simulation.set_physical_parameters(physical_params)
@@ -47,11 +77,11 @@ def run_simulation(physical_params=default_physical_params,
simulation.set_runtime_parameters(runtime_params)
simulation.set_writer(writer)
- no_slip_bc = NoSlipBoundaryCondition()
+ no_slip_bc = cpu.boundaryconditions.NoSlipBoundaryCondition()
block_thickness = 3 * grid_params.node_distance
simulation.add_object(
- GbCuboid3D(
+ cpu.geometry.GbCuboid3D(
grid_params.bounding_box.min_x1 - block_thickness,
grid_params.bounding_box.min_x2 - block_thickness,
grid_params.bounding_box.min_x3 - block_thickness,
@@ -59,10 +89,10 @@ def run_simulation(physical_params=default_physical_params,
grid_params.bounding_box.max_x2 + block_thickness,
grid_params.bounding_box.min_x3),
no_slip_bc,
- State.SOLID, "/geo/addWallZMin")
+ cpu.geometry.State.SOLID, "/geo/addWallZMin")
simulation.add_object(
- GbCuboid3D(
+ cpu.geometry.GbCuboid3D(
grid_params.bounding_box.min_x1 - block_thickness,
grid_params.bounding_box.min_x2 - block_thickness,
grid_params.bounding_box.max_x3,
@@ -70,7 +100,7 @@ def run_simulation(physical_params=default_physical_params,
grid_params.bounding_box.max_x2 + block_thickness,
grid_params.bounding_box.max_x3 + block_thickness),
no_slip_bc,
- State.SOLID, "/geo/addWallZMax")
+ cpu.geometry.State.SOLID, "/geo/addWallZMax")
simulation.run_simulation()
diff --git a/Python/poiseuille/test_poiseuille_l2.py b/Python/poiseuille/test_poiseuille_l2.py
index 93aa2600d5260dea7e72f3aa98db7334fe5285c6..818cba40e115945c60e4fa2ac96b3b6b5ab0bba8 100644
--- a/Python/poiseuille/test_poiseuille_l2.py
+++ b/Python/poiseuille/test_poiseuille_l2.py
@@ -1,3 +1,37 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file test_poiseuille_l2.py
+! \ingroup poiseuille
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
import os
import shutil
import unittest
@@ -5,8 +39,7 @@ import unittest
import matplotlib.pyplot as plt
import numpy as np
import pyvista as pv
-from pyfluids.cpu.kernel import LBMKernel, KernelType
-from pyfluids.cpu.parameters import GridParameters, PhysicalParameters, RuntimeParameters
+from pyfluids import cpu
from scipy import stats
from errors import normalized_l2_error
@@ -33,13 +66,13 @@ class TestPoiseuilleFlow(unittest.TestCase):
self.skipTest("This test is not implemented correctly yet")
plt.ion()
- physical_params = PhysicalParameters()
+ physical_params = cpu.parameters.PhysicalParameters()
- runtime_params = RuntimeParameters()
+ runtime_params = cpu.parameters.RuntimeParameters()
runtime_params.number_of_threads = os.cpu_count()
runtime_params.timestep_log_interval = 10000
- kernel = LBMKernel(KernelType.CompressibleCumulantFourthOrderViscosity)
+ kernel = cpu.kernel.LBMKernel(cpu.kernel.KernelType.CompressibleCumulantFourthOrderViscosity)
kernel.use_forcing = True
normalized_l2_errors = []
@@ -140,7 +173,7 @@ def get_heights_from_indices(mesh, indices):
def create_grid_params_with_nodes_in_column(nodes_in_column, delta_x):
- grid_params = GridParameters()
+ grid_params = cpu.parameters.GridParameters()
grid_params.node_distance = delta_x
grid_params.number_of_nodes_per_direction = [1, 1, nodes_in_column]
grid_params.blocks_per_direction = [1, 1, 8]
diff --git a/Python/tests/test_acousticscaling.py b/Python/tests/test_acousticscaling.py
index 6413123a80db8c5882fcf1dbe6f72a1f5438736c..02454b935e3a147e045f45c273392646aeca6b8c 100644
--- a/Python/tests/test_acousticscaling.py
+++ b/Python/tests/test_acousticscaling.py
@@ -1,9 +1,41 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file test_acousticscaling.py
+! \ingroup tests
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
import unittest
from typing import List
-from pyfluids.cpu.kernel import LBMKernel, KernelType
-from pyfluids.cpu.parameters import GridParameters, PhysicalParameters, RuntimeParameters
-
+from pyfluids import cpu
from acousticscaling import OneDirectionalAcousticScaling
@@ -58,18 +90,18 @@ class OneDirectionalAcousticScalingTest(unittest.TestCase):
self.assertEqual(self.grid_params.periodic_boundary_in_x2, actual_grid_params.periodic_boundary_in_x2)
self.assertEqual(self.grid_params.periodic_boundary_in_x3, actual_grid_params.periodic_boundary_in_x3)
- def assert_physical_params_scaled_by_factor(self, actual_params: PhysicalParameters, factor: int):
+ def assert_physical_params_scaled_by_factor(self, actual_params: cpu.parameters.PhysicalParameters, factor: int):
self.assertEqual(self.physical_params.lattice_viscosity * factor, actual_params.lattice_viscosity)
self.assertEqual(self.physical_params.bulk_viscosity_factor, actual_params.bulk_viscosity_factor)
- def assert_runtime_params_scaled_by_factor(self, actual_params: RuntimeParameters, factor: int):
+ def assert_runtime_params_scaled_by_factor(self, actual_params: cpu.parameters.RuntimeParameters, factor: int):
self.assertEqual(self.runtime_params.number_of_timesteps * factor, actual_params.number_of_timesteps)
self.assertEqual(self.runtime_params.number_of_threads, actual_params.number_of_threads)
self.assertEqual(self.runtime_params.timestep_log_interval, actual_params.timestep_log_interval)
- def assert_kernel_forcing_scaled_by_factor(self, actual_kernel: LBMKernel, factor: int):
+ def assert_kernel_forcing_scaled_by_factor(self, actual_kernel: cpu.kernel.LBMKernel, factor: int):
self.assertEqual(self.kernel.type, actual_kernel.type)
- self.assertEqual(self.kernel.use_forcing, actual_kernel.use_forcing)
+ self.assertEqual(self.kernel.use_forcing, actual_kernel.cpu.parameters.use_forcing)
self.assertAlmostEqual(self.kernel.forcing_in_x1 / factor, actual_kernel.forcing_in_x1)
self.assertAlmostEqual(self.kernel.forcing_in_x2, actual_kernel.forcing_in_x2)
self.assertAlmostEqual(self.kernel.forcing_in_x3, actual_kernel.forcing_in_x3)
@@ -80,14 +112,14 @@ class OneDirectionalAcousticScalingTest(unittest.TestCase):
@staticmethod
def make_kernel():
- kernel = LBMKernel(KernelType.CompressibleCumulantFourthOrderViscosity)
+ kernel = cpu.kernel.LBMKernel(cpu.kernel.KernelType.CompressibleCumulantFourthOrderViscosity)
kernel.use_forcing = True
kernel.forcing_in_x1 = 5e-10
return kernel
@staticmethod
def make_runtime_params():
- runtime_params = RuntimeParameters()
+ runtime_params = cpu.parameters.RuntimeParameters()
runtime_params.number_of_threads = 4
runtime_params.number_of_timesteps = 4_000_000
runtime_params.timestep_log_interval = 1_000_000
@@ -95,13 +127,13 @@ class OneDirectionalAcousticScalingTest(unittest.TestCase):
@staticmethod
def make_physical_params():
- physical_params = PhysicalParameters()
+ physical_params = cpu.parameters.PhysicalParameters()
physical_params.lattice_viscosity = 1e-4
return physical_params
@staticmethod
def make_grid_params():
- grid_params = GridParameters()
+ grid_params = cpu.parameters.GridParameters()
grid_params.node_distance = 1
grid_params.number_of_nodes_per_direction = [1, 1, 16]
grid_params.blocks_per_direction = [1, 1, 16]
diff --git a/Python/tests/test_boundaryconditions.py b/Python/tests/test_boundaryconditions.py
index e004ddfa21c78ea3d63a89f5dbc3bd7438a18ff1..d914c50cad2051188331b2efe604907091fa731e 100644
--- a/Python/tests/test_boundaryconditions.py
+++ b/Python/tests/test_boundaryconditions.py
@@ -1,5 +1,39 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file test_boundaryconditions.py
+! \ingroup tests
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
import unittest
-from pyfluids.cpu.boundaryconditions import *
+from pyfluids import cpu
class BoundaryConditionsTest(unittest.TestCase):
@@ -8,13 +42,13 @@ class BoundaryConditionsTest(unittest.TestCase):
"""
Should be able to create NoSlipBoundaryCondition
"""
- sut = NoSlipBoundaryCondition()
+ sut = cpu.boundaryconditions.NoSlipBoundaryCondition()
def test__can_create_velocity_bc(self):
"""
Should be able to create VelocityBoundaryCondition
"""
- sut = VelocityBoundaryCondition()
+ sut = cpu.boundaryconditions.VelocityBoundaryCondition()
def test__can_create_velocity_bc_with_directions_function_and_time(self):
"""
@@ -24,7 +58,7 @@ class BoundaryConditionsTest(unittest.TestCase):
parser = Parser()
parser.expression = "1"
- sut = VelocityBoundaryCondition(True, True, True, parser, 0, 1)
+ sut = cpu.boundaryconditions.VelocityBoundaryCondition(True, True, True, parser, 0, 1)
def test__can_create_velocity_bc_with_directions__function_per_direction__and__time(self):
"""
@@ -40,7 +74,7 @@ class BoundaryConditionsTest(unittest.TestCase):
f3 = Parser()
f3.expression = "1"
- sut = VelocityBoundaryCondition(True, True, True, f1, f2, f3, 0, 1)
+ sut = cpu.boundaryconditions.VelocityBoundaryCondition(True, True, True, f1, f2, f3, 0, 1)
def test__can_create_velocity_bc_with_speeds_and_times_per_direction(self):
"""
@@ -51,11 +85,11 @@ class BoundaryConditionsTest(unittest.TestCase):
start2, end2 = 1, 2
start3, end3 = 2, 3
- sut = VelocityBoundaryCondition(vx1, start1, end1, vx2, start2, end2, vx3, start3, end3)
+ sut = cpu.boundaryconditions.VelocityBoundaryCondition(vx1, start1, end1, vx2, start2, end2, vx3, start3, end3)
def test__can_create_non_reflecting_outflow(self):
"""
Should be able to create NonReflectingOutflow
"""
- sut = NonReflectingOutflow()
+ sut = cpu.boundaryconditions.NonReflectingOutflow()
diff --git a/Python/tests/test_geometry.py b/Python/tests/test_geometry.py
index 5bb89eb245b6055653b78fde381da050d402b0cc..3d297f5c176cd99f7969adf37333588d86b77627 100644
--- a/Python/tests/test_geometry.py
+++ b/Python/tests/test_geometry.py
@@ -1,6 +1,40 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file test_geometry.py
+! \ingroup tests
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
import unittest
-from pyfluids.cpu.geometry import *
+from pyfluids import cpu
class TestGeometry(unittest.TestCase):
@@ -9,7 +43,7 @@ class TestGeometry(unittest.TestCase):
"""
WHEN setting point coordinates in constructor THEN point should have coordinates
"""
- sut = GbPoint3D(4, 8, 3)
+ sut = cpu.geometry.GbPoint3D(4, 8, 3)
self.assertEqual(sut.x1, 4)
self.assertEqual(sut.x2, 8)
@@ -19,7 +53,7 @@ class TestGeometry(unittest.TestCase):
"""
WHEN setting point coordinates THEN point should have coordinates
"""
- sut = GbPoint3D()
+ sut = cpu.geometry.GbPoint3D()
sut.x1 = 4
sut.x2 = 8
@@ -33,10 +67,10 @@ class TestGeometry(unittest.TestCase):
"""
WHEN setting line points THEN line should have points
"""
- sut = GbLine3D()
+ sut = cpu.geometry.GbLine3D()
- point1 = GbPoint3D()
- point2 = GbPoint3D()
+ point1 = cpu.geometry.GbPoint3D()
+ point2 = cpu.geometry.GbPoint3D()
sut.point1 = point1
sut.point2 = point2
diff --git a/Python/tests/test_kernel.py b/Python/tests/test_kernel.py
index 8f58a1c869f9e292856268d43245a75f1dcfe213..e0159bec6802cb08d73214038b177091879fee46 100644
--- a/Python/tests/test_kernel.py
+++ b/Python/tests/test_kernel.py
@@ -1,12 +1,46 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file test_kernel.py
+! \ingroup tests
+! \author Sven Marcus, Henry Korb
+=======================================================================================
+"""
import unittest
-from pyfluids.cpu.kernel import LBMKernel, KernelType
+from pyfluids import cpu
class TestLBMKernel(unittest.TestCase):
def setUp(self) -> None:
- self.sut = LBMKernel(KernelType.BGK)
+ self.sut = cpu.kernel.LBMKernel(cpu.kernel.KernelType.BGK)
def test_lbm_kernel__when_use_forcing_set_to_true__use_forcing_should_be_true(self) -> None:
"""
@@ -57,4 +91,4 @@ class TestLBMKernel(unittest.TestCase):
"""
actual = self.sut.type
- self.assertEqual(KernelType.BGK, actual)
+ self.assertEqual(cpu.kernel.KernelType.BGK, actual)
diff --git a/apps/gpu/LBM/ActuatorLine/ActuatorLine.cpp b/apps/gpu/LBM/ActuatorLine/ActuatorLine.cpp
index 58e5aede18b9c4197b4d21b129c6347023b9390e..9d982ebac0059b4512041194100f6e1fdfa61924 100644
--- a/apps/gpu/LBM/ActuatorLine/ActuatorLine.cpp
+++ b/apps/gpu/LBM/ActuatorLine/ActuatorLine.cpp
@@ -1,4 +1,35 @@
-
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file ActuatorLine.cpp
+//! \ingroup ActuatorLine
+//! \author Henry Korb, Henrik Asmuth
+//=======================================================================================
#define _USE_MATH_DEFINES
#include <math.h>
#include <string>
@@ -28,12 +59,14 @@
#include "GridGenerator/grid/GridBuilder/LevelGridBuilder.h"
#include "GridGenerator/grid/GridBuilder/MultipleGridBuilder.h"
#include "GridGenerator/grid/BoundaryConditions/Side.h"
+#include "GridGenerator/grid/BoundaryConditions/BoundaryCondition.h"
+
#include "GridGenerator/grid/GridFactory.h"
#include "GridGenerator/io/SimulationFileWriter/SimulationFileWriter.h"
#include "GridGenerator/io/GridVTKWriter/GridVTKWriter.h"
-#include "GridGenerator/io/STLReaderWriter/STLReader.h"
-#include "GridGenerator/io/STLReaderWriter/STLWriter.h"
+#include "GridGenerator/TransientBCSetter/TransientBCSetter.h"
+
//////////////////////////////////////////////////////////////////////////
@@ -44,10 +77,12 @@
#include "VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.h"
#include "VirtualFluids_GPU/Parameter/Parameter.h"
#include "VirtualFluids_GPU/Output/FileWriter.h"
-#include "VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.h"
+#include "VirtualFluids_GPU/PreCollisionInteractor/ActuatorFarm.h"
#include "VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.h"
#include "VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.h"
#include "VirtualFluids_GPU/Factories/BoundaryConditionFactory.h"
+#include "VirtualFluids_GPU/TurbulenceModels/TurbulenceModelFactory.h"
+#include "VirtualFluids_GPU/Factories/GridScalingFactory.h"
#include "VirtualFluids_GPU/GPU/CudaMemoryManager.h"
@@ -63,26 +98,10 @@
LbmOrGks lbmOrGks = LBM;
-const real reference_diameter = 126.0; // diameter in m
-
-const real L_x = 10*reference_diameter;
-const real L_y = 6*reference_diameter;
-const real L_z = 6*reference_diameter;
-
-const real viscosity = 1.56e-5;
-
-const real velocity = 9.0;
-
-const real mach = 0.1;
-
-const uint nodes_per_diameter = 16;
-
std::string path(".");
std::string simulationName("ActuatorLine");
-const float tOut = 100;
-const float tEnd = 280; // total time of simulation in s
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
@@ -98,30 +117,59 @@ void multipleLevel(const std::string& configPath)
vf::gpu::Communicator& communicator = vf::gpu::Communicator::getInstance();
auto gridFactory = GridFactory::make();
- gridFactory->setTriangularMeshDiscretizationMethod(TriangularMeshDiscretizationMethod::POINT_IN_OBJECT);
auto gridBuilder = MultipleGridBuilder::makeShared(gridFactory);
- ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ vf::basics::ConfigurationFile config;
+ config.load(configPath);
+
+ const real reference_diameter = config.getValue<real>("ReferenceDiameter");
+ const uint nodes_per_diameter = config.getValue<uint>("NodesPerDiameter");
+ const real velocity = config.getValue<real>("Velocity");
+
+
+ const real L_x = 24*reference_diameter;
+ const real L_y = 6*reference_diameter;
+ const real L_z = 6*reference_diameter;
+
+ const real viscosity = 1.56e-5;
+
+ const real mach = 0.1;
+
+
+ const float tStartOut = config.getValue<real>("tStartOut");
+ const float tOut = config.getValue<real>("tOut");
+ const float tEnd = config.getValue<real>("tEnd"); // total time of simulation
+
+ const float tStartAveraging = config.getValue<real>("tStartAveraging");
+ const float tStartTmpAveraging = config.getValue<real>("tStartTmpAveraging");
+ const float tAveraging = config.getValue<real>("tAveraging");
+ const float tStartOutProbe = config.getValue<real>("tStartOutProbe");
+ const float tOutProbe = config.getValue<real>("tOutProbe");
+
+ SPtr<Parameter> para = std::make_shared<Parameter>(communicator.getNummberOfProcess(), communicator.getPID(), &config);
+ BoundaryConditionFactory bcFactory = BoundaryConditionFactory();
+ GridScalingFactory scalingFactory = GridScalingFactory();
+
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
const real dx = reference_diameter/real(nodes_per_diameter);
+ real turbPos[3] = {3*reference_diameter, 3*reference_diameter, 3*reference_diameter};
+
gridBuilder->addCoarseGrid(0.0, 0.0, 0.0,
L_x, L_y, L_z, dx);
+ gridBuilder->setNumberOfLayers(4,0);
+ gridBuilder->addGrid( new Cuboid( turbPos[0]-1.5*reference_diameter, turbPos[1]-1.5*reference_diameter, turbPos[2]-1.5*reference_diameter,
+ turbPos[0]+10.0*reference_diameter, turbPos[1]+1.5*reference_diameter, turbPos[2]+1.5*reference_diameter) , 1 );
+ para->setMaxLevel(2);
+ scalingFactory.setScalingFactory(GridScalingFactory::GridScaling::ScaleCompressible);
+
gridBuilder->setPeriodicBoundaryCondition(false, false, false);
gridBuilder->buildGrids(lbmOrGks, false); // buildGrids() has to be called before setting the BCs!!!!
- ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-
- vf::basics::ConfigurationFile config;
- config.load(configPath);
- ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////^
- SPtr<Parameter> para = std::make_shared<Parameter>(communicator.getNummberOfProcess(), communicator.getPID(), &config);
- BoundaryConditionFactory bcFactory = BoundaryConditionFactory();
- ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
const real dt = dx * mach / (sqrt(3) * velocity);
@@ -140,14 +188,11 @@ void multipleLevel(const std::string& configPath)
para->setPrintFiles(true);
- para->setMaxLevel(1);
-
-
para->setVelocityLB(velocityLB);
para->setViscosityLB(viscosityLB);
para->setVelocityRatio( dx / dt );
para->setViscosityRatio( dx*dx/dt );
- para->setMainKernel("CumulantK17CompChim");
+ para->setMainKernel("CumulantK17");
para->setInitialCondition([&](real coordX, real coordY, real coordZ, real &rho, real &vx, real &vy, real &vz) {
rho = (real)0.0;
@@ -156,13 +201,15 @@ void multipleLevel(const std::string& configPath)
vz = (real)0.0;
});
+ para->setTimestepStartOut( uint(tStartOut/dt) );
para->setTimestepOut( uint(tOut/dt) );
para->setTimestepEnd( uint(tEnd/dt) );
para->setIsBodyForce( true );
-
+ para->setUseStreams( true );
/////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+
gridBuilder->setVelocityBoundaryCondition(SideType::MX, velocityLB, 0.0, 0.0);
gridBuilder->setVelocityBoundaryCondition(SideType::MY, velocityLB, 0.0, 0.0);
@@ -172,42 +219,52 @@ void multipleLevel(const std::string& configPath)
gridBuilder->setPressureBoundaryCondition(SideType::PX, 0.0);
bcFactory.setVelocityBoundaryCondition(BoundaryConditionFactory::VelocityBC::VelocityAndPressureCompressible);
- bcFactory.setPressureBoundaryCondition(BoundaryConditionFactory::PressureBC::PressureNonEquilibriumCompressible);
+ bcFactory.setPressureBoundaryCondition(BoundaryConditionFactory::PressureBC::OutflowNonReflective);
+
+ SPtr<TurbulenceModelFactory> tmFactory = std::make_shared<TurbulenceModelFactory>(para);
+ tmFactory->readConfigFile(config);
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- real turbPos[3] = {3*reference_diameter, 3*reference_diameter, 3*reference_diameter};
- real epsilon = 5.f; // width of gaussian smearing
- real density = 1.225f;
- int level = 0;
- uint nBlades = 3;
- uint nBladeNodes = 32;
+ int level = 1; // grid level at which the turbine samples velocities and distributes forces
+ const real epsilon = dx*exp2(-level)*1.5; // width of gaussian smearing
+ const real density = 1.225f;
+ const uint nBlades = 3;
+ const uint nBladeNodes = 32;
+ const real tipspeed_ratio = 7.5f; // tipspeed ratio = angular vel * radius / inflow vel
+ const real omega = 2*tipspeed_ratio*velocity/reference_diameter;
+
+
+ SPtr<ActuatorFarm> actuator_farm = std::make_shared<ActuatorFarm>(nBlades, density, nBladeNodes, epsilon, level, dt, dx, true);
+ std::vector<real> bladeRadii;
+ real dr = reference_diameter/(nBladeNodes*2);
+ for(uint node=0; node<nBladeNodes; node++){ bladeRadii.emplace_back(dr*(node+1)); }
+ actuator_farm->addTurbine(turbPos[0], turbPos[1], turbPos[2], reference_diameter, omega, 0, 0, bladeRadii);
+ para->addActuator( actuator_farm );
- SPtr<ActuatorLine> actuator_line =SPtr<ActuatorLine>( new ActuatorLine(nBlades, density, nBladeNodes, epsilon, turbPos[0], turbPos[1], turbPos[2], reference_diameter, level, dt, dx) );
- para->addActuator( actuator_line );
- SPtr<PointProbe> pointProbe = SPtr<PointProbe>( new PointProbe("pointProbe", para->getOutputPath(), 100, 1, 500, 100) );
- std::vector<real> probeCoordsX = {reference_diameter,2*reference_diameter,5*reference_diameter};
- std::vector<real> probeCoordsY = {3*reference_diameter,3*reference_diameter,3*reference_diameter};
- std::vector<real> probeCoordsZ = {3*reference_diameter,3*reference_diameter,3*reference_diameter};
- pointProbe->addProbePointsFromList(probeCoordsX, probeCoordsY, probeCoordsZ);
- // pointProbe->addProbePointsFromXNormalPlane(2*D, 0.0, 0.0, L_y, L_z, (uint)L_y/dx, (uint)L_z/dx);
+ // SPtr<PointProbe> pointProbe = std::make_shared<PointProbe>("pointProbe", para->getOutputPath(), 100, 1, 500, 100);
+ // std::vector<real> probeCoordsX = {reference_diameter,2*reference_diameter,5*reference_diameter};
+ // std::vector<real> probeCoordsY = {3*reference_diameter,3*reference_diameter,3*reference_diameter};
+ // std::vector<real> probeCoordsZ = {3*reference_diameter,3*reference_diameter,3*reference_diameter};
+ // pointProbe->addProbePointsFromList(probeCoordsX, probeCoordsY, probeCoordsZ);
+ // // pointProbe->addProbePointsFromXNormalPlane(2*D, 0.0, 0.0, L_y, L_z, (uint)L_y/dx, (uint)L_z/dx);
- pointProbe->addStatistic(Statistic::Means);
- pointProbe->addStatistic(Statistic::Variances);
- para->addProbe( pointProbe );
+ // pointProbe->addStatistic(Statistic::Means);
+ // pointProbe->addStatistic(Statistic::Variances);
+ // para->addProbe( pointProbe );
- SPtr<PlaneProbe> planeProbe = SPtr<PlaneProbe>( new PlaneProbe("planeProbe", para->getOutputPath(), 100, 500, 100, 100) );
- planeProbe->setProbePlane(5*reference_diameter, 0, 0, dx, L_y, L_z);
- planeProbe->addStatistic(Statistic::Means);
- para->addProbe( planeProbe );
+ // SPtr<PlaneProbe> planeProbe = std::make_shared<PlaneProbe>("planeProbe", para->getOutputPath(), 100, 500, 100, 100);
+ // planeProbe->setProbePlane(5*reference_diameter, 0, 0, dx, L_y, L_z);
+ // planeProbe->addStatistic(Statistic::Means);
+ // para->addProbe( planeProbe );
auto cudaMemoryManager = std::make_shared<CudaMemoryManager>(para);
auto gridGenerator = GridProvider::makeGridGenerator(gridBuilder, para, cudaMemoryManager, communicator);
- Simulation sim(para, cudaMemoryManager, communicator, *gridGenerator, &bcFactory);
+ Simulation sim(para, cudaMemoryManager, communicator, *gridGenerator, &bcFactory, tmFactory, &scalingFactory);
sim.run();
}
diff --git a/apps/gpu/LBM/ActuatorLine/configActuatorLine.txt b/apps/gpu/LBM/ActuatorLine/configActuatorLine.txt
index 233994f0d32a48190d84f7044500e24b06b926a9..5799f24716777295b2f835ab00561ff767ba87b9 100644
--- a/apps/gpu/LBM/ActuatorLine/configActuatorLine.txt
+++ b/apps/gpu/LBM/ActuatorLine/configActuatorLine.txt
@@ -6,3 +6,29 @@ Path = .
#informations for reading
##################################################
GridPath=.
+##################################################
+ReferenceDiameter=126
+NodesPerDiameter=32
+Velocity=9
+##################################################
+tStartOut=100
+tOut=100
+tEnd=1000
+##################################################
+
+tStartTmpAveraging=100
+tStartAveraging=100
+tAveraging=100
+tTmpAveraging=100
+tStartOutProbe=100
+tOutProbe=100
+
+##################################################
+#TurbulenceModel = QR
+#SGSconstant = 0.3333333
+#
+#QuadricLimiterP = 100000.0
+#QuadricLimiterM = 100000.0
+#QuadricLimiterD = 100000.0
+##################################################
+
diff --git a/apps/gpu/LBM/BoundaryLayer/BoundaryLayer.cpp b/apps/gpu/LBM/BoundaryLayer/BoundaryLayer.cpp
index 991025b649d69305c030fe2f1dd1763a2137af9b..5fc31904433bfe2df0722ab1c63f574d3fcb9a35 100644
--- a/apps/gpu/LBM/BoundaryLayer/BoundaryLayer.cpp
+++ b/apps/gpu/LBM/BoundaryLayer/BoundaryLayer.cpp
@@ -1,4 +1,35 @@
-
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file BoundaryLayer.cpp
+//! \ingroup BoundaryLayer
+//! \author Henry Korb, Henrik Asmuth
+//=======================================================================================
#define _USE_MATH_DEFINES
#include <math.h>
#include <string>
@@ -8,6 +39,7 @@
#include <fstream>
#include <exception>
#include <memory>
+#include <numeric>
//////////////////////////////////////////////////////////////////////////
@@ -19,6 +51,7 @@
#include "Core/VectorTypes.h"
#include <basics/config/ConfigurationFile.h>
+#include "lbm/constants/NumericConstants.h"
#include <logger/Logger.h>
@@ -28,12 +61,16 @@
#include "GridGenerator/grid/GridBuilder/LevelGridBuilder.h"
#include "GridGenerator/grid/GridBuilder/MultipleGridBuilder.h"
#include "GridGenerator/grid/BoundaryConditions/Side.h"
+#include "GridGenerator/grid/BoundaryConditions/BoundaryCondition.h"
+
#include "GridGenerator/grid/GridFactory.h"
+#include "geometries/Cuboid/Cuboid.h"
+#include "geometries/TriangularMesh/TriangularMesh.h"
+
#include "GridGenerator/io/SimulationFileWriter/SimulationFileWriter.h"
#include "GridGenerator/io/GridVTKWriter/GridVTKWriter.h"
-#include "GridGenerator/io/STLReaderWriter/STLReader.h"
-#include "GridGenerator/io/STLReaderWriter/STLWriter.h"
+#include "GridGenerator/TransientBCSetter/TransientBCSetter.h"
//////////////////////////////////////////////////////////////////////////
@@ -44,24 +81,28 @@
#include "VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.h"
#include "VirtualFluids_GPU/Parameter/Parameter.h"
#include "VirtualFluids_GPU/Output/FileWriter.h"
-#include "VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.h"
#include "VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.h"
#include "VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.h"
#include "VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.h"
#include "VirtualFluids_GPU/PreCollisionInteractor/Probes/WallModelProbe.h"
+#include "VirtualFluids_GPU/PreCollisionInteractor/PrecursorWriter.h"
#include "VirtualFluids_GPU/Factories/BoundaryConditionFactory.h"
+#include "VirtualFluids_GPU/Factories/GridScalingFactory.h"
#include "VirtualFluids_GPU/TurbulenceModels/TurbulenceModelFactory.h"
#include "VirtualFluids_GPU/GPU/CudaMemoryManager.h"
+#include "utilities/communication.h"
+
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
std::string path(".");
-std::string simulationName("BoundayLayer");
+std::string simulationName("BoundaryLayer");
+using namespace vf::lbm::constant;
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
@@ -87,8 +128,16 @@ void multipleLevel(const std::string& configPath)
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////^
SPtr<Parameter> para = std::make_shared<Parameter>(communicator.getNummberOfProcess(), communicator.getPID(), &config);
BoundaryConditionFactory bcFactory = BoundaryConditionFactory();
-
+ GridScalingFactory scalingFactory = GridScalingFactory();
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+
+ const int nProcs = communicator.getNummberOfProcess();
+ const uint procID = vf::gpu::Communicator::getInstance().getPID();
+ std::vector<uint> devices(10);
+ std::iota(devices.begin(), devices.end(), 0);
+ para->setDevices(devices);
+ para->setMaxDev(nProcs);
+
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
//
@@ -100,23 +149,45 @@ void multipleLevel(const std::string& configPath)
LbmOrGks lbmOrGks = LBM;
- const real H = 1000.0; // boundary layer height in m
+ const real H = config.getValue("boundaryLayerHeight", 1000.0); // boundary layer height in m
const real L_x = 6*H;
const real L_y = 4*H;
- const real L_z = 1*H;
+ const real L_z = H;
+
+ const real z0 = config.getValue("z0", 0.1f); // roughness length in m
+ const real u_star = config.getValue("u_star", 0.4f); //friction velocity in m/s
+ const real kappa = config.getValue("vonKarmanConstant", 0.4f); // von Karman constant
- const real z0 = 0.1; // roughness length in m
- const real u_star = 0.4; //friction velocity in m/s
- const real kappa = 0.4; // von Karman constant
+ const real viscosity = config.getValue("viscosity", 1.56e-5f);
- const real viscosity = 1.56e-5;
+ const real velocity = 0.5f*u_star/kappa*log(H/z0+1.f); //0.5 times max mean velocity at the top in m/s
- const real velocity = 0.5*u_star/kappa*log(L_z/z0); //0.5 times max mean velocity at the top in m/s
+ const real mach = config.getValue<real>("Ma", 0.1);
- const real mach = config.contains("Ma")? config.getValue<real>("Ma"): 0.1;
+ const uint nodes_per_H = config.getValue<uint>("nz", 64);
- const uint nodes_per_H = config.contains("nz")? config.getValue<uint>("nz"): 64;
+ const bool writePrecursor = config.getValue("writePrecursor", false);
+ bool useDistributions;
+ std::string precursorDirectory;
+ int nTWritePrecursor; real tStartPrecursor, posXPrecursor;
+ if(writePrecursor)
+ {
+ nTWritePrecursor = config.getValue<int>("nTimestepsWritePrecursor");
+ tStartPrecursor = config.getValue<real>("tStartPrecursor");
+ posXPrecursor = config.getValue<real>("posXPrecursor");
+ useDistributions = config.getValue<bool>("useDistributions", false);
+ precursorDirectory = config.getValue<std::string>("precursorDirectory");
+ }
+
+ const bool readPrecursor = config.getValue("readPrecursor", false);
+ int timestepsBetweenReadsPrecursor;
+ if(readPrecursor)
+ {
+ timestepsBetweenReadsPrecursor = config.getValue<int>("nTimestepsReadPrecursor");
+ precursorDirectory = config.getValue<std::string>("precursorDirectory");
+ useDistributions = config.getValue<bool>("useDistributions", false);
+ }
// all in s
const float tStartOut = config.getValue<real>("tStartOut");
@@ -130,7 +201,7 @@ void multipleLevel(const std::string& configPath)
const float tOutProbe = config.getValue<real>("tOutProbe");
- const real dx = L_z/real(nodes_per_H);
+ const real dx = H/real(nodes_per_H);
const real dt = dx * mach / (sqrt(3) * velocity);
@@ -155,15 +226,17 @@ void multipleLevel(const std::string& configPath)
para->setPrintFiles(true);
- para->setForcing(pressureGradientLB, 0, 0);
+ if(!readPrecursor) para->setForcing(pressureGradientLB, 0, 0);
para->setVelocityLB(velocityLB);
para->setViscosityLB(viscosityLB);
para->setVelocityRatio( dx / dt );
para->setViscosityRatio( dx*dx/dt );
para->setDensityRatio( 1.0 );
- para->setMainKernel("TurbulentViscosityCumulantK17CompChim");
-
+ bool useStreams = (nProcs > 1 ? true: false);
+ // useStreams=false;
+ para->setUseStreams(useStreams);
+ para->setMainKernel("CumulantK17");
para->setIsBodyForce( config.getValue<bool>("bodyForce") );
para->setTimestepStartOut(uint(tStartOut/dt) );
@@ -172,64 +245,206 @@ void multipleLevel(const std::string& configPath)
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- SPtr<TurbulenceModelFactory> tmFactory = SPtr<TurbulenceModelFactory>( new TurbulenceModelFactory(para) );
+ SPtr<TurbulenceModelFactory> tmFactory = std::make_shared<TurbulenceModelFactory>(para);
tmFactory->readConfigFile( config );
+ /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+
+ const real xSplit = L_x/nProcs;
+ const real overlap = 8.0*dx;
+
+ real xMin = procID * xSplit;
+ real xMax = (procID+1) * xSplit;
+ real xGridMin = procID * xSplit;
+ real xGridMax = (procID+1) * xSplit;
+
+ real yMin = 0.0;
+ real yMax = L_y;
+ real zMin = 0.0;
+ real zMax = L_z;
+
+ bool isFirstSubDomain = (procID == 0 && nProcs > 1)? true: false;
+ bool isLastSubDomain = (procID == nProcs-1 && nProcs > 1)? true: false;
+ bool isMidSubDomain = (!isFirstSubDomain && !isLastSubDomain && nProcs > 1)? true: false;
- // tmFactory->setTurbulenceModel(TurbulenceModel::AMD);
- // tmFactory->setModelConstant(config.getValue<real>("SGSconstant"));
+ if(isFirstSubDomain)
+ {
+ xGridMax += overlap;
+ if(!readPrecursor) xGridMin -= overlap;
+ }
+ if(isLastSubDomain)
+ {
+ xGridMin -= overlap;
+ if(!readPrecursor) xGridMax += overlap;
+ }
+ if(isMidSubDomain)
+ {
+ xGridMax += overlap;
+ xGridMin -= overlap;
+ }
- /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ gridBuilder->addCoarseGrid( xGridMin, 0.0, 0.0,
+ xGridMax, L_y, L_z, dx);
+ if(true)// Add refinement
+ {
+ gridBuilder->setNumberOfLayers(4,0);
+ real xMaxRefinement = readPrecursor? xGridMax-H: xGridMax; //Stop refinement some distance before outlet if domain ist not periodic
+ gridBuilder->addGrid( new Cuboid( xGridMin, 0.f, 0.f, xMaxRefinement, L_y, 0.5*L_z) , 1 );
+ para->setMaxLevel(2);
+ scalingFactory.setScalingFactory(GridScalingFactory::GridScaling::ScaleCompressible);
+ }
+
+ if(nProcs > 1)
+ {
+ gridBuilder->setSubDomainBox(
+ std::make_shared<BoundingBox>(xMin, xMax, yMin, yMax, zMin, zMax));
+ gridBuilder->setPeriodicBoundaryCondition(false, true, false);
+ }
+ else
+ {
+ gridBuilder->setPeriodicBoundaryCondition(!readPrecursor, true, false);
+ }
- gridBuilder->addCoarseGrid(0.0, 0.0, 0.0,
- L_x, L_y, L_z, dx);
- // gridBuilder->setNumberOfLayers(12, 8);
+ gridBuilder->buildGrids(lbmOrGks, true); // buildGrids() has to be called before setting the BCs!!!!
- // gridBuilder->addGrid( new Cuboid( 0.0, 0.0, 0.0, L_x, L_y, 0.3*L_z) , 1 );
- // para->setMaxLevel(2);
+ std::cout << "nProcs: "<< nProcs << "Proc: " << procID << " isFirstSubDomain: " << isFirstSubDomain << " isLastSubDomain: " << isLastSubDomain << " isMidSubDomain: " << isMidSubDomain << std::endl;
+
+ if(nProcs > 1){
+ if (isFirstSubDomain || isMidSubDomain) {
+ gridBuilder->findCommunicationIndices(CommunicationDirections::PX, lbmOrGks);
+ gridBuilder->setCommunicationProcess(CommunicationDirections::PX, procID+1);
+ }
- gridBuilder->setPeriodicBoundaryCondition(true, true, false);
+ if (isLastSubDomain || isMidSubDomain) {
+ gridBuilder->findCommunicationIndices(CommunicationDirections::MX, lbmOrGks);
+ gridBuilder->setCommunicationProcess(CommunicationDirections::MX, procID-1);
+ }
- gridBuilder->buildGrids(lbmOrGks, false); // buildGrids() has to be called before setting the BCs!!!!
+ if (isFirstSubDomain && !readPrecursor) {
+ gridBuilder->findCommunicationIndices(CommunicationDirections::MX, lbmOrGks);
+ gridBuilder->setCommunicationProcess(CommunicationDirections::MX, nProcs-1);
+ }
+ if (isLastSubDomain && !readPrecursor) {
+ gridBuilder->findCommunicationIndices(CommunicationDirections::PX, lbmOrGks);
+ gridBuilder->setCommunicationProcess(CommunicationDirections::PX, 0);
+ }
+ }
uint samplingOffset = 2;
- // gridBuilder->setVelocityBoundaryCondition(SideType::MZ, 0.0, 0.0, 0.0);
+
+ std::cout << " precursorDirectory " << precursorDirectory << std::endl;
+
+ if(readPrecursor)
+ {
+ if(isFirstSubDomain || nProcs == 1)
+ {
+ auto precursor = createFileCollection(precursorDirectory + "/precursor", FileType::VTK);
+ gridBuilder->setPrecursorBoundaryCondition(SideType::MX, precursor, timestepsBetweenReadsPrecursor);
+ // gridBuilder->setVelocityBoundaryCondition(SideType::MX, velocityLB, 0.0, 0.0);
+ }
+
+ if(isLastSubDomain || nProcs == 1)
+ {
+ gridBuilder->setPressureBoundaryCondition(SideType::PX, 0.f);
+ }
+ }
+
gridBuilder->setStressBoundaryCondition(SideType::MZ,
0.0, 0.0, 1.0, // wall normals
- samplingOffset, z0/dx); // wall model settinng
- para->setHasWallModelMonitor(true);
- bcFactory.setStressBoundaryCondition(BoundaryConditionFactory::StressBC::StressPressureBounceBack);
+ samplingOffset, z0, dx); // wall model settinng
+ para->setHasWallModelMonitor(true);
+ gridBuilder->setSlipBoundaryCondition(SideType::PZ, 0.0f, 0.0f, -1.0f);
- gridBuilder->setSlipBoundaryCondition(SideType::PZ, 0.0, 0.0, 0.0);
+ bcFactory.setVelocityBoundaryCondition(BoundaryConditionFactory::VelocityBC::VelocityCompressible);
+ bcFactory.setStressBoundaryCondition(BoundaryConditionFactory::StressBC::StressPressureBounceBack);
bcFactory.setSlipBoundaryCondition(BoundaryConditionFactory::SlipBC::SlipBounceBack);
-
+ bcFactory.setPressureBoundaryCondition(BoundaryConditionFactory::PressureBC::OutflowNonReflective);
+ bcFactory.setPrecursorBoundaryCondition(useDistributions ? BoundaryConditionFactory::PrecursorBC::DistributionsPrecursor : BoundaryConditionFactory::PrecursorBC::VelocityPrecursor);
+ para->setOutflowPressureCorrectionFactor(0.0);
- real cPi = 3.1415926535897932384626433832795;
- para->setInitialCondition([&](real coordX, real coordY, real coordZ, real &rho, real &vx, real &vy, real &vz) {
+ if(readPrecursor)
+ {
+ para->setInitialCondition([&](real coordX, real coordY, real coordZ, real &rho, real &vx, real &vy, real &vz) {
rho = (real)0.0;
- vx = (u_star/0.4 * log(coordZ/z0) + 2.0*sin(cPi*16.0f*coordX/L_x)*sin(cPi*8.0f*coordZ/H)/(pow(coordZ/H,c2o1)+c1o1)) * dt / dx;
- vy = 2.0*sin(cPi*16.0f*coordX/L_x)*sin(cPi*8.0f*coordZ/H)/(pow(coordZ/H,c2o1)+c1o1) * dt / dx;
- vz = 8.0*u_star/0.4*(sin(cPi*8.0*coordY/H)*sin(cPi*8.0*coordZ/H)+sin(cPi*8.0*coordX/L_x))/(pow(L_z/2.0-coordZ, c2o1)+c1o1) * dt / dx;
- });
- ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ vx = rho = c0o1;
+ vx = u_star/c4o10*(u_star/c4o10 * log(coordZ/z0+c1o1)) * dt/dx;
+ vy = c0o1;
+ vz = c0o1;
+ });
+ }
+ else
+ {
+ para->setInitialCondition([&](real coordX, real coordY, real coordZ, real &rho, real &vx, real &vy, real &vz) {
+ rho = (real)0.0;
+ vx = rho = c0o1;
+ vx = (u_star/c4o10 * log(coordZ/z0+c1o1) + c2o1*sin(cPi*c16o1*coordX/L_x)*sin(cPi*c8o1*coordZ/H)/(pow(coordZ/H,c2o1)+c1o1)) * dt/dx;
+ vy = c2o1*sin(cPi*c16o1*coordX/L_x)*sin(cPi*c8o1*coordZ/H)/(pow(coordZ/H,c2o1)+c1o1) * dt/dx;
+ vz = c8o1*u_star/c4o10*(sin(cPi*c8o1*coordY/H)*sin(cPi*c8o1*coordZ/H)+sin(cPi*c8o1*coordX/L_x))/(pow(c1o2*L_z-coordZ, c2o1)+c1o1) * dt/dx;
+ });
+ }
+
+
- SPtr<PlanarAverageProbe> planarAverageProbe = SPtr<PlanarAverageProbe>( new PlanarAverageProbe("planeProbe", para->getOutputPath(), tStartAveraging/dt, tStartTmpAveraging/dt, tAveraging/dt , tStartOutProbe/dt, tOutProbe/dt, 'z') );
- planarAverageProbe->addAllAvailableStatistics();
- planarAverageProbe->setFileNameToNOut();
- para->addProbe( planarAverageProbe );
+ //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ if(!readPrecursor && (isFirstSubDomain || nProcs == 1))
+ {
+ SPtr<PlanarAverageProbe> planarAverageProbe = SPtr<PlanarAverageProbe>( new PlanarAverageProbe("planeProbe", para->getOutputPath(), tStartAveraging/dt, tStartTmpAveraging/dt, tAveraging/dt , tStartOutProbe/dt, tOutProbe/dt, 'z') );
+ planarAverageProbe->addAllAvailableStatistics();
+ planarAverageProbe->setFileNameToNOut();
+ para->addProbe( planarAverageProbe );
+
+ para->setHasWallModelMonitor(true);
+ SPtr<WallModelProbe> wallModelProbe = SPtr<WallModelProbe>( new WallModelProbe("wallModelProbe", para->getOutputPath(), tStartAveraging/dt, tStartTmpAveraging/dt, tAveraging/dt/4.0 , tStartOutProbe/dt, tOutProbe/dt) );
+ wallModelProbe->addAllAvailableStatistics();
+ wallModelProbe->setFileNameToNOut();
+ wallModelProbe->setForceOutputToStress(true);
+ if(para->getIsBodyForce())
+ wallModelProbe->setEvaluatePressureGradient(true);
+ para->addProbe( wallModelProbe );
+ }
+
+ SPtr<PlaneProbe> planeProbe1 = SPtr<PlaneProbe>( new PlaneProbe("planeProbe_1", para->getOutputPath(), tStartAveraging/dt, 10, tStartOutProbe/dt, tOutProbe/dt) );
+ planeProbe1->setProbePlane(100.0, 0.0, 0, dx, L_y, L_z);
+ planeProbe1->addAllAvailableStatistics();
+ para->addProbe( planeProbe1 );
+
+ if(readPrecursor)
+ {
+ SPtr<PlaneProbe> planeProbe2 = SPtr<PlaneProbe>( new PlaneProbe("planeProbe_2", para->getOutputPath(), tStartAveraging/dt, 10, tStartOutProbe/dt, tOutProbe/dt) );
+ planeProbe2->setProbePlane(1000.0, 0.0, 0, dx, L_y, L_z);
+ planeProbe2->addAllAvailableStatistics();
+ para->addProbe( planeProbe2 );
+
+ SPtr<PlaneProbe> planeProbe3 = SPtr<PlaneProbe>( new PlaneProbe("planeProbe_3", para->getOutputPath(), tStartAveraging/dt, 10, tStartOutProbe/dt, tOutProbe/dt) );
+ planeProbe3->setProbePlane(1500.0, 0.0, 0, dx, L_y, L_z);
+ planeProbe3->addAllAvailableStatistics();
+ para->addProbe( planeProbe3 );
+
+ SPtr<PlaneProbe> planeProbe4 = SPtr<PlaneProbe>( new PlaneProbe("planeProbe_4", para->getOutputPath(), tStartAveraging/dt, 10, tStartOutProbe/dt, tOutProbe/dt) );
+ planeProbe4->setProbePlane(2000.0, 0.0, 0, dx, L_y, L_z);
+ planeProbe4->addAllAvailableStatistics();
+ para->addProbe( planeProbe4 );
+
+ SPtr<PlaneProbe> planeProbe5 = SPtr<PlaneProbe>( new PlaneProbe("planeProbe_5", para->getOutputPath(), tStartAveraging/dt, 10, tStartOutProbe/dt, tOutProbe/dt) );
+ planeProbe5->setProbePlane(2500.0, 0.0, 0, dx, L_y, L_z);
+ planeProbe5->addAllAvailableStatistics();
+ para->addProbe( planeProbe5 );
+
+ SPtr<PlaneProbe> planeProbe6 = SPtr<PlaneProbe>( new PlaneProbe("planeProbe_6", para->getOutputPath(), tStartAveraging/dt, 10, tStartOutProbe/dt, tOutProbe/dt) );
+ planeProbe6->setProbePlane(0.0, L_y/2.0, 0, L_x, dx, L_z);
+ planeProbe6->addAllAvailableStatistics();
+ para->addProbe( planeProbe6 );
+ }
- para->setHasWallModelMonitor(true);
- SPtr<WallModelProbe> wallModelProbe = SPtr<WallModelProbe>( new WallModelProbe("wallModelProbe", para->getOutputPath(), tStartAveraging/dt, tStartTmpAveraging/dt, tAveraging/dt/4.0 , tStartOutProbe/dt, tOutProbe/dt) );
- wallModelProbe->addAllAvailableStatistics();
- wallModelProbe->setFileNameToNOut();
- wallModelProbe->setForceOutputToStress(true);
- if(para->getIsBodyForce())
- wallModelProbe->setEvaluatePressureGradient(true);
- para->addProbe( wallModelProbe );
+ if(writePrecursor)
+ {
+ SPtr<PrecursorWriter> precursorWriter = std::make_shared<PrecursorWriter>("precursor", para->getOutputPath()+precursorDirectory, posXPrecursor, 0, L_y, 0, L_z, tStartPrecursor/dt, nTWritePrecursor, useDistributions? OutputVariable::Distributions: OutputVariable::Velocities, 1000);
+ para->addProbe(precursorWriter);
+ }
auto cudaMemoryManager = std::make_shared<CudaMemoryManager>(para);
auto gridGenerator = GridProvider::makeGridGenerator(gridBuilder, para, cudaMemoryManager, communicator);
- Simulation sim(para, cudaMemoryManager, communicator, *gridGenerator, &bcFactory, tmFactory);
+ Simulation sim(para, cudaMemoryManager, communicator, *gridGenerator, &bcFactory, tmFactory, &scalingFactory);
sim.run();
}
diff --git a/apps/gpu/LBM/BoundaryLayer/configBoundaryLayer.txt b/apps/gpu/LBM/BoundaryLayer/configBoundaryLayer.txt
index a489f0ab89738a193b16fee41c212a5943f6525d..83e7861a5fb85ea800d187699f1c6c1409422f0a 100644
--- a/apps/gpu/LBM/BoundaryLayer/configBoundaryLayer.txt
+++ b/apps/gpu/LBM/BoundaryLayer/configBoundaryLayer.txt
@@ -7,7 +7,7 @@ Path = .
##################################################
GridPath = .
##################################################
-Devices = 1
+Devices = 0
##################################################
tStartOut = 0
tOut = 100000
@@ -28,3 +28,15 @@ SGSconstant = 0.2
QuadricLimiterP = 100000.0
QuadricLimiterM = 100000.0
QuadricLimiterD = 100000.0
+
+##################################################
+readPrecursor = false
+nTimestepsReadPrecursor = 10
+precursorFile = precursor/Precursor
+
+##################################################
+writePrecursor = false
+nTimestepsWritePrecursor = 10
+
+tStartPrecursor = 100
+posXPrecursor = 3000
\ No newline at end of file
diff --git a/apps/gpu/LBM/TGV_3D_MultiGPU/TGV_3D_MultiGPU.cpp b/apps/gpu/LBM/TGV_3D_MultiGPU/TGV_3D_MultiGPU.cpp
index 8c303dc07c911c363e892ce53f7bfe7f48e284d6..045c208274bc6bc216d25e8c2fa905916a52f87b 100644
--- a/apps/gpu/LBM/TGV_3D_MultiGPU/TGV_3D_MultiGPU.cpp
+++ b/apps/gpu/LBM/TGV_3D_MultiGPU/TGV_3D_MultiGPU.cpp
@@ -1,7 +1,38 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file TGV_3d_MuitiGPU.cpp
+//! \ingroup TGV_3D_MultiGPU
+//! \author Martin Schoenherr
+//=======================================================================================
//#define MPI_LOGGING
//Martin Branch
-
#include <mpi.h>
#if defined( MPI_LOGGING )
#include <mpe.h>
@@ -97,7 +128,7 @@ bool useWale = false;
int mpirank;
int mpiWorldSize;
-std::string kernel( "CumulantK20Comp" );
+std::string kernel( "CumulantK17CompChim" );
//std::string path("F:/Work/Computations/out/TaylorGreen3DNew/"); //LEGOLAS
//std::string path("results/"); //PHOENIX
diff --git a/pyproject.toml b/pyproject.toml
index 8fcb7926102d188b44d8c74084235b6f175edf80..257da6fd95d683081dbff865c864079eae9c675d 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,2 +1,9 @@
[build-system]
-requires = ["setuptools", "wheel", "scikit-build"]
\ No newline at end of file
+requires = [
+ "setuptools>=42",
+ "scikit-build",
+ "cmake",
+ "ninja; platform_system!='Windows'"
+]
+build-backend = "setup_builder"
+backend-path = ["utilities"]
\ No newline at end of file
diff --git a/pythonbindings/CMakeLists.txt b/pythonbindings/CMakeLists.txt
index 5a84adef027fdfa2953e016693bb64570e48c1ef..644b308c97760d6a8243ae7782f0737a1d168100 100644
--- a/pythonbindings/CMakeLists.txt
+++ b/pythonbindings/CMakeLists.txt
@@ -1,24 +1,35 @@
project(VirtualFluidsPython LANGUAGES CUDA CXX)
+pybind11_add_module(python_bindings MODULE src/VirtualFluids.cpp)
+
+set_target_properties( python_bindings PROPERTIES
+ LIBRARY_OUTPUT_DIRECTORY ${CMAKE_SOURCE_DIR}/pythonbindings/pyfluids
+ OUTPUT_NAME "bindings")
+
IF(BUILD_VF_GPU)
- pybind11_add_module(pyfluids src/VirtualFluidsModulesGPU.cpp)
- set_source_files_properties(src/VirtualFluidsModulesGPU.cpp PROPERTIES LANGUAGE CUDA)
+ set_source_files_properties(src/VirtualFluids.cpp PROPERTIES LANGUAGE CUDA)
- target_link_libraries(pyfluids PRIVATE GridGenerator VirtualFluids_GPU basics lbmCuda logger)
- target_include_directories(pyfluids PRIVATE ${VF_THIRD_DIR}/cuda_samples/)
+ target_include_directories(python_bindings PRIVATE ${VF_THIRD_DIR}/cuda_samples/)
+ target_compile_definitions(python_bindings PRIVATE VF_GPU_PYTHONBINDINGS)
+ target_link_libraries(python_bindings PRIVATE GridGenerator VirtualFluids_GPU basics lbmCuda logger)
ENDIF()
+
IF(BUILD_VF_CPU)
- pybind11_add_module(pyfluids src/VirtualFluidsModulesCPU.cpp)
- pybind11_add_module(pymuparser src/muParser.cpp)
+ target_compile_definitions(python_bindings PRIVATE VF_METIS VF_MPI VF_CPU_PYTHONBINDINGS)
+ target_link_libraries(python_bindings PRIVATE simulationconfig VirtualFluidsCore muparser basics)
+
+ # include bindings for muparsers
+ pybind11_add_module(pymuparser MODULE src/muParser.cpp)
# TODO: Move this to MuParser CMakeLists.txt
set_target_properties(muparser PROPERTIES POSITION_INDEPENDENT_CODE ON)
- target_compile_definitions(pyfluids PRIVATE VF_METIS VF_MPI)
+ set_target_properties( pymuparser PROPERTIES
+ LIBRARY_OUTPUT_DIRECTORY ${CMAKE_SOURCE_DIR}/pythonbindings/pymuparser
+ OUTPUT_NAME "bindings")
target_compile_definitions(pymuparser PRIVATE VF_METIS VF_MPI)
-
- target_link_libraries(pyfluids PRIVATE simulationconfig VirtualFluidsCore muparser basics)
target_link_libraries(pymuparser PRIVATE muparser)
ENDIF()
-target_include_directories(pyfluids PRIVATE ${CMAKE_SOURCE_DIR}/src/)
-target_include_directories(pyfluids PRIVATE ${CMAKE_BINARY_DIR})
\ No newline at end of file
+
+target_include_directories(python_bindings PRIVATE ${CMAKE_SOURCE_DIR}/src/)
+target_include_directories(python_bindings PRIVATE ${CMAKE_BINARY_DIR})
\ No newline at end of file
diff --git a/Python/boundary_layer/__init__.py b/pythonbindings/pyfluids-stubs/__init__.pyi
similarity index 100%
rename from Python/boundary_layer/__init__.py
rename to pythonbindings/pyfluids-stubs/__init__.pyi
diff --git a/pythonbindings/pyfluids-stubs/bindings/__init__.pyi b/pythonbindings/pyfluids-stubs/bindings/__init__.pyi
new file mode 100644
index 0000000000000000000000000000000000000000..4e7f353eab97cc536f8f18e72319af1cd7a1916a
--- /dev/null
+++ b/pythonbindings/pyfluids-stubs/bindings/__init__.pyi
@@ -0,0 +1,38 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file __init__.pyi
+! \ingroup bindings
+! \author Henry Korb
+=======================================================================================
+"""
+class ostream_redirect:
+ def __init__(self, stdout: bool = ..., stderr: bool = ...) -> None: ...
+ def __enter__(self) -> None: ...
+ def __exit__(self, *args) -> None: ...
diff --git a/pythonbindings/pyfluids-stubs/bindings/basics/__init__.pyi b/pythonbindings/pyfluids-stubs/bindings/basics/__init__.pyi
new file mode 100644
index 0000000000000000000000000000000000000000..a41b7934ca706dc0db5bd6188fee3150456e0cd9
--- /dev/null
+++ b/pythonbindings/pyfluids-stubs/bindings/basics/__init__.pyi
@@ -0,0 +1,82 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file __init__.py
+! \ingroup basics
+! \author Henry Korb
+=======================================================================================
+"""
+from typing import ClassVar
+
+from typing import overload
+
+class ConfigurationFile:
+ def __init__(self) -> None: ...
+ def contains(self, key: str) -> bool: ...
+ @overload
+ def get_bool_value(self, key: str) -> bool: ...
+ @overload
+ def get_bool_value(self, key: str, default_value: bool) -> bool: ...
+ @overload
+ def get_double_value(self, key: str) -> float: ...
+ @overload
+ def get_double_value(self, key: str, default_value: float) -> float: ...
+ @overload
+ def get_float_value(self, key: str) -> float: ...
+ @overload
+ def get_float_value(self, key: str, default_value: float) -> float: ...
+ @overload
+ def get_int_value(self, key: str) -> int: ...
+ @overload
+ def get_int_value(self, key: str, default_value: int) -> int: ...
+ @overload
+ def get_string_value(self, key: str) -> str: ...
+ @overload
+ def get_string_value(self, key: str, default_value: str) -> str: ...
+ @overload
+ def get_uint_value(self, key: str) -> int: ...
+ @overload
+ def get_uint_value(self, key: str, default_value: int) -> int: ...
+ def load(self, file: str) -> bool: ...
+
+class LbmOrGks:
+ __members__: ClassVar[dict] = ... # read-only
+ GKS: ClassVar[LbmOrGks] = ...
+ LBM: ClassVar[LbmOrGks] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
diff --git a/pythonbindings/pyfluids-stubs/bindings/basics/logger.pyi b/pythonbindings/pyfluids-stubs/bindings/basics/logger.pyi
new file mode 100644
index 0000000000000000000000000000000000000000..43938ff7646efd3c596ae29971cce39fed865fa6
--- /dev/null
+++ b/pythonbindings/pyfluids-stubs/bindings/basics/logger.pyi
@@ -0,0 +1,83 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file logger.pyi
+! \ingroup basics
+! \author Henry Korb
+=======================================================================================
+"""
+from typing import Any, ClassVar
+
+log: None
+
+class Level:
+ __members__: ClassVar[dict] = ... # read-only
+ INFO_HIGH: ClassVar[Level] = ...
+ INFO_INTERMEDIATE: ClassVar[Level] = ...
+ INFO_LOW: ClassVar[Level] = ...
+ LOGGER_ERROR: ClassVar[Level] = ...
+ WARNING: ClassVar[Level] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class Logger:
+ def __init__(self, *args, **kwargs) -> None: ...
+ @staticmethod
+ def add_stdout() -> None: ...
+ @staticmethod
+ def enable_printed_rank_numbers(print: bool) -> None: ...
+ @staticmethod
+ def set_debug_level(level: int) -> None: ...
+ @staticmethod
+ def time_stamp(time_stemp: TimeStamp) -> None: ...
+
+class TimeStamp:
+ __members__: ClassVar[dict] = ... # read-only
+ DISABLE: ClassVar[TimeStamp] = ...
+ ENABLE: ClassVar[TimeStamp] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
diff --git a/pythonbindings/pyfluids-stubs/bindings/gpu/__init__.pyi b/pythonbindings/pyfluids-stubs/bindings/gpu/__init__.pyi
new file mode 100644
index 0000000000000000000000000000000000000000..36c2fea76713e980bb95eb6726d778de8c9a6583
--- /dev/null
+++ b/pythonbindings/pyfluids-stubs/bindings/gpu/__init__.pyi
@@ -0,0 +1,436 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file __init__.pyi
+! \ingroup gpu
+! \author Henry Korb
+=======================================================================================
+"""
+from typing import Any, Callable, ClassVar, List, Optional
+
+from typing import overload
+import numpy
+import pyfluids.bindings.basics
+import pyfluids.bindings.gpu.grid_generator as grid_generator
+
+class ActuatorFarm(PreCollisionInteractor):
+ def __init__(self, number_of_blades_per_turbine: int, density: float, number_of_nodes_per_blade: int, epsilon: float, level: int, delta_t: float, delta_x: float, use_host_arrays: bool) -> None: ...
+ def add_turbine(self, posX: float, posY: float, posZ: float, diameter: float, omega: float, azimuth: float, yaw: float, bladeRadii: List[float]) -> None: ...
+ def calc_blade_forces(self) -> None: ...
+ def get_all_azimuths(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_coords_x(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_coords_x_device(self) -> int: ...
+ def get_all_blade_coords_y(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_coords_y_device(self) -> int: ...
+ def get_all_blade_coords_z(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_coords_z_device(self) -> int: ...
+ def get_all_blade_forces_x(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_forces_x_device(self) -> int: ...
+ def get_all_blade_forces_y(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_forces_y_device(self) -> int: ...
+ def get_all_blade_forces_z(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_forces_z_device(self) -> int: ...
+ def get_all_blade_radii(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_radii_device(self) -> int: ...
+ def get_all_blade_velocities_x(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_velocities_x_device(self) -> int: ...
+ def get_all_blade_velocities_y(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_velocities_y_device(self) -> int: ...
+ def get_all_blade_velocities_z(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_blade_velocities_z_device(self) -> int: ...
+ def get_all_omegas(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_turbine_pos_x(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_turbine_pos_y(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_turbine_pos_z(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_all_yaws(self) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_azimuth(self, turbine: int) -> float: ...
+ def get_turbine_blade_coords_x(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_coords_x_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_coords_y(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_coords_y_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_coords_z(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_coords_z_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_forces_x(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_forces_x_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_forces_y(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_forces_y_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_forces_z(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_forces_z_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_radii(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_radii_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_velocities_x(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_velocities_x_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_velocities_y(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_velocities_y_device(self, turbine: int) -> int: ...
+ def get_turbine_blade_velocities_z(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_blade_velocities_z_device(self, turbine: int) -> int: ...
+ def get_turbine_omega(self, turbine: int) -> float: ...
+ def get_turbine_pos(self, turbine: int) -> numpy.ndarray[numpy.float32]: ...
+ def get_turbine_yaw(self, turbine: int) -> float: ...
+ def set_all_azimuths(self, azimuths: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_all_blade_coords(self, blade_coords_x: numpy.ndarray[numpy.float32], blade_coords_y: numpy.ndarray[numpy.float32], blade_coords_z: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_all_blade_forces(self, blade_forces_x: numpy.ndarray[numpy.float32], blade_forces_y: numpy.ndarray[numpy.float32], blade_forces_z: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_all_blade_velocities(self, blade_velocities_x: numpy.ndarray[numpy.float32], blade_velocities_y: numpy.ndarray[numpy.float32], blade_velocities_z: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_all_omegas(self, omegas: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_all_yaws(self, yaws: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_turbine_azimuth(self, turbine: int, azimuth: float) -> None: ...
+ def set_turbine_blade_coords(self, turbine: int, blade_coords_x: numpy.ndarray[numpy.float32], blade_coords_y: numpy.ndarray[numpy.float32], blade_coords_z: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_turbine_blade_forces(self, turbine: int, blade_forces_x: numpy.ndarray[numpy.float32], blade_forces_y: numpy.ndarray[numpy.float32], blade_forces_z: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_turbine_blade_velocities(self, turbine: int, blade_velocities_x: numpy.ndarray[numpy.float32], blade_velocities_y: numpy.ndarray[numpy.float32], blade_velocities_z: numpy.ndarray[numpy.float32]) -> None: ...
+ def set_turbine_omega(self, turbine: int, omega: float) -> None: ...
+ def set_turbine_yaw(self, turbine: int, yaw: float) -> None: ...
+ @property
+ def delta_t(self) -> float: ...
+ @property
+ def delta_x(self) -> float: ...
+ @property
+ def density(self) -> float: ...
+ @property
+ def number_of_blades_per_turbine(self) -> int: ...
+ @property
+ def number_of_indices(self) -> int: ...
+ @property
+ def number_of_nodes(self) -> int: ...
+ @property
+ def number_of_nodes_per_blade(self) -> int: ...
+ @property
+ def number_of_turbines(self) -> int: ...
+
+class BoundaryConditionFactory:
+ def __init__(self) -> None: ...
+ def set_geometry_boundary_condition(self, boundary_condition_type) -> None: ...
+ def set_no_slip_boundary_condition(self, boundary_condition_type) -> None: ...
+ def set_precursor_boundary_condition(self, boundary_condition_type) -> None: ...
+ def set_pressure_boundary_condition(self, boundary_condition_type) -> None: ...
+ def set_slip_boundary_condition(self, boundary_condition_type) -> None: ...
+ def set_stress_boundary_condition(self, boundary_condition_type) -> None: ...
+ def set_velocity_boundary_condition(self, boundary_condition_type) -> None: ...
+
+class Communicator:
+ def __init__(self, *args, **kwargs) -> None: ...
+ @staticmethod
+ def get_instance() -> Communicator: ...
+ def get_number_of_process(self) -> int: ...
+ def get_pid(self) -> int: ...
+
+class CudaMemoryManager:
+ def __init__(self, parameter: Parameter) -> None: ...
+
+class FileType:
+ __members__: ClassVar[dict] = ... # read-only
+ VTK: ClassVar[FileType] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class GridProvider:
+ def __init__(self, *args, **kwargs) -> None: ...
+ @staticmethod
+ def make_grid_generator(builder: grid_generator.GridBuilder, para: Parameter, cuda_memory_manager: CudaMemoryManager, communicator: Communicator) -> GridProvider: ...
+
+class GridScaling:
+ __members__: ClassVar[dict] = ... # read-only
+ NotSpecified: ClassVar[GridScaling] = ...
+ ScaleCompressible: ClassVar[GridScaling] = ...
+ ScaleRhoSq: ClassVar[GridScaling] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class GridScalingFactory:
+ def __init__(self) -> None: ...
+ def set_scaling_factory(self, scaling_type) -> None: ...
+
+class NoSlipBC:
+ __members__: ClassVar[dict] = ... # read-only
+ NoSlip3rdMomentsCompressible: ClassVar[NoSlipBC] = ...
+ NoSlipBounceBack: ClassVar[NoSlipBC] = ...
+ NoSlipCompressible: ClassVar[NoSlipBC] = ...
+ NoSlipImplicitBounceBack: ClassVar[NoSlipBC] = ...
+ NoSlipIncompressible: ClassVar[NoSlipBC] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class OutputVariable:
+ __members__: ClassVar[dict] = ... # read-only
+ Distributions: ClassVar[OutputVariable] = ...
+ Velocities: ClassVar[OutputVariable] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class Parameter:
+ @overload
+ def __init__(self, number_of_processes: int, my_ID: int, config_data: Optional[pyfluids.bindings.basics.ConfigurationFile]) -> None: ...
+ @overload
+ def __init__(self, number_of_processes: int, my_ID: int) -> None: ...
+ @overload
+ def __init__(self, config_data: pyfluids.bindings.basics.ConfigurationFile) -> None: ...
+ def add_actuator(self, actuator: PreCollisionInteractor) -> None: ...
+ def add_probe(self, probe: PreCollisionInteractor) -> None: ...
+ def get_SGS_constant(self) -> float: ...
+ def get_density_ratio(self) -> float: ...
+ def get_force_ratio(self) -> float: ...
+ def get_is_body_force(self) -> bool: ...
+ def get_output_path(self) -> str: ...
+ def get_output_prefix(self) -> str: ...
+ def get_velocity(self) -> float: ...
+ def get_velocity_ratio(self) -> float: ...
+ def get_viscosity(self) -> float: ...
+ def get_viscosity_ratio(self) -> float: ...
+ def set_AD_kernel(self, ad_kernel: str) -> None: ...
+ def set_calc_turbulence_intensity(self, calc_velocity_and_fluctuations: bool) -> None: ...
+ def set_comp_on(self, is_comp: bool) -> None: ...
+ def set_density_ratio(self, density_ratio: float) -> None: ...
+ def set_devices(self, devices: List[int]) -> None: ...
+ def set_diff_on(self, is_diff: bool) -> None: ...
+ def set_forcing(self, forcing_x: float, forcing_y: float, forcing_z: float) -> None: ...
+ def set_has_wall_model_monitor(self, has_wall_monitor: bool) -> None: ...
+ def set_initial_condition(self, init_func: Callable[[float,float,float],List[float]]) -> None: ...
+ def set_initial_condition_log_law(self, u_star: float, z0: float, velocity_ratio: float) -> None: ...
+ def set_initial_condition_perturbed_log_law(self, u_star: float, z0: float, length_x: float, length_z: float, height: float, velocity_ratio: float) -> None: ...
+ def set_initial_condition_uniform(self, velocity_x: float, velocity_y: float, velocity_z: float) -> None: ...
+ def set_is_body_force(self, is_body_force: bool) -> None: ...
+ def set_main_kernel(self, kernel: str) -> None: ...
+ def set_max_dev(self, max_dev: int) -> None: ...
+ def set_max_level(self, number_of_levels: int) -> None: ...
+ def set_outflow_pressure_correction_factor(self, correction_factor: float) -> None: ...
+ def set_output_path(self, o_path: str) -> None: ...
+ def set_output_prefix(self, o_prefix: str) -> None: ...
+ def set_print_files(self, print_files: bool) -> None: ...
+ def set_quadric_limiters(self, quadric_limiter_p: float, quadric_limiter_m: float, quadric_limiter_d: float) -> None: ...
+ def set_temperature_BC(self, temp_bc: float) -> None: ...
+ def set_temperature_init(self, temp: float) -> None: ...
+ def set_timestep_end(self, tend: int) -> None: ...
+ def set_timestep_of_coarse_level(self, timestep: int) -> None: ...
+ def set_timestep_out(self, tout: int) -> None: ...
+ def set_timestep_start_out(self, t_start_out: int) -> None: ...
+ def set_use_streams(self, use_streams: bool) -> None: ...
+ def set_velocity_LB(self, velocity: float) -> None: ...
+ def set_velocity_ratio(self, velocity_ratio: float) -> None: ...
+ def set_viscosity_LB(self, viscosity: float) -> None: ...
+ def set_viscosity_ratio(self, viscosity_ratio: float) -> None: ...
+
+class PreCollisionInteractor:
+ def __init__(self, *args, **kwargs) -> None: ...
+
+class PrecursorBC:
+ __members__: ClassVar[dict] = ... # read-only
+ DistributionsPrecursor: ClassVar[PrecursorBC] = ...
+ NotSpecified: ClassVar[PrecursorBC] = ...
+ VelocityPrecursor: ClassVar[PrecursorBC] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class PrecursorWriter(PreCollisionInteractor):
+ def __init__(self, filename: str, output_path: str, x_pos: float, y_min: float, y_max: float, z_min: float, z_max: float, t_start_out: int, t_save: int, output_variable: OutputVariable, max_timesteps_per_file: int) -> None: ...
+
+class PressureBC:
+ __members__: ClassVar[dict] = ... # read-only
+ NotSpecified: ClassVar[PressureBC] = ...
+ OutflowNonReflective: ClassVar[PressureBC] = ...
+ OutflowNonReflectivePressureCorrection: ClassVar[PressureBC] = ...
+ PressureEquilibrium: ClassVar[PressureBC] = ...
+ PressureEquilibrium2: ClassVar[PressureBC] = ...
+ PressureNonEquilibriumCompressible: ClassVar[PressureBC] = ...
+ PressureNonEquilibriumIncompressible: ClassVar[PressureBC] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class SideType:
+ __members__: ClassVar[dict] = ... # read-only
+ GEOMETRY: ClassVar[SideType] = ...
+ MX: ClassVar[SideType] = ...
+ MY: ClassVar[SideType] = ...
+ MZ: ClassVar[SideType] = ...
+ PX: ClassVar[SideType] = ...
+ PY: ClassVar[SideType] = ...
+ PZ: ClassVar[SideType] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class Simulation:
+ @overload
+ def __init__(self, parameter: Parameter, memoryManager: CudaMemoryManager, communicator, gridProvider: GridProvider, bcFactory: BoundaryConditionFactory, gridScalingFactory: GridScalingFactory) -> None: ...
+ @overload
+ def __init__(self, parameter: Parameter, memoryManager: CudaMemoryManager, communicator, gridProvider: GridProvider, bcFactory: BoundaryConditionFactory) -> None: ...
+ @overload
+ def __init__(self, parameter: Parameter, memoryManager: CudaMemoryManager, communicator, gridProvider: GridProvider, bcFactory: BoundaryConditionFactory, tmFactory: TurbulenceModelFactory, gridScalingFactory: GridScalingFactory) -> None: ...
+ def addEnstrophyAnalyzer(self, t_analyse: int) -> None: ...
+ def addKineticEnergyAnalyzer(self, t_analyse: int) -> None: ...
+ def run(self) -> None: ...
+
+class SlipBC:
+ __members__: ClassVar[dict] = ... # read-only
+ NotSpecified: ClassVar[SlipBC] = ...
+ SlipBounceBack: ClassVar[SlipBC] = ...
+ SlipCompressible: ClassVar[SlipBC] = ...
+ SlipCompressibleTurbulentViscosity: ClassVar[SlipBC] = ...
+ SlipIncompressible: ClassVar[SlipBC] = ...
+ SlipPressureCompressibleTurbulentViscosity: ClassVar[SlipBC] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class StressBC:
+ __members__: ClassVar[dict] = ... # read-only
+ NotSpecified: ClassVar[StressBC] = ...
+ StressBounceBack: ClassVar[StressBC] = ...
+ StressCompressible: ClassVar[StressBC] = ...
+ StressPressureBounceBack: ClassVar[StressBC] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class TurbulenceModel:
+ __members__: ClassVar[dict] = ... # read-only
+ AMD: ClassVar[TurbulenceModel] = ...
+ NONE: ClassVar[TurbulenceModel] = ...
+ QR: ClassVar[TurbulenceModel] = ...
+ Smagorinsky: ClassVar[TurbulenceModel] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class TurbulenceModelFactory:
+ def __init__(self, para: Parameter) -> None: ...
+ def read_config_file(self, config_data: pyfluids.bindings.basics.ConfigurationFile) -> None: ...
+ def set_model_constant(self, model_constant: float) -> None: ...
+ def set_turbulence_model(self, turbulence_model: TurbulenceModel) -> None: ...
+
+class VTKFileCollection(FileCollection):
+ def __init__(self, prefix: str) -> None: ...
+
+class VelocityBC:
+ __members__: ClassVar[dict] = ... # read-only
+ NotSpecified: ClassVar[VelocityBC] = ...
+ VelocityAndPressureCompressible: ClassVar[VelocityBC] = ...
+ VelocityCompressible: ClassVar[VelocityBC] = ...
+ VelocityIncompressible: ClassVar[VelocityBC] = ...
+ VelocitySimpleBounceBackCompressible: ClassVar[VelocityBC] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class FileCollection:
+ def __init__(self, *args, **kwargs) -> None: ...
+
+def create_file_collection(prefix: str, type: FileType) -> FileCollection: ...
diff --git a/pythonbindings/pyfluids-stubs/bindings/gpu/grid_generator.pyi b/pythonbindings/pyfluids-stubs/bindings/gpu/grid_generator.pyi
new file mode 100644
index 0000000000000000000000000000000000000000..8d715e4b4cd49e6dbf92da3aedddbc4b869067c4
--- /dev/null
+++ b/pythonbindings/pyfluids-stubs/bindings/gpu/grid_generator.pyi
@@ -0,0 +1,100 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file grid_generator.pyi
+! \ingroup gpu
+! \author Henry Korb
+=======================================================================================
+"""
+from typing import Any, List
+
+from typing import overload
+import pyfluids.bindings.basics
+import pyfluids.bindings.gpu
+
+class BoundingBox:
+ def __init__(self, min_x: float, max_x: float, min_y: float, max_y: float, min_z: float, max_z: float) -> None: ...
+
+class Conglomerate(Object):
+ def __init__(self, *args, **kwargs) -> None: ...
+ def add(self, object: Object) -> None: ...
+ @staticmethod
+ def make_shared() -> Conglomerate: ...
+ def subtract(self, object: Object) -> None: ...
+
+class Cuboid(Object):
+ def __init__(self, min_x1: float, min_x2: float, min_x3: float, max_x1: float, max_x2: float, max_x3: float) -> None: ...
+
+class GridBuilder:
+ def __init__(self, *args, **kwargs) -> None: ...
+ def get_number_of_grid_levels(self) -> int: ...
+
+class GridFactory:
+ def __init__(self, *args, **kwargs) -> None: ...
+ @staticmethod
+ def make() -> GridFactory: ...
+
+class LevelGridBuilder(GridBuilder):
+ def __init__(self, *args, **kwargs) -> None: ...
+ def set_no_slip_boundary_condition(self, side_type: pyfluids.bindings.gpu.SideType) -> None: ...
+ def set_periodic_boundary_condition(self, periodic_x: bool, periodic_y: bool, periodic_z: bool) -> None: ...
+ def set_precursor_boundary_condition(self, side_type: pyfluids.bindings.gpu.SideType, file_collection: pyfluids.bindings.gpu.VelocityFileCollection, n_t_read: int, velocity_x: float = ..., velocity_y: float = ..., velocity_z: float = ..., file_level_to_grid_level_map: List[int] = ...) -> None: ...
+ def set_pressure_boundary_condition(self, side_type: pyfluids.bindings.gpu.SideType, rho: float) -> None: ...
+ def set_slip_boundary_condition(self, side_type: pyfluids.bindings.gpu.SideType, normal_x: float, normal_y: float, normal_z: float) -> None: ...
+ def set_stress_boundary_condition(self, side_type: pyfluids.bindings.gpu.SideType, normal_x: float, normal_y: float, normal_z: float, sampling_offset: int, z0: float, dx: float) -> None: ...
+ def set_velocity_boundary_condition(self, side_type: pyfluids.bindings.gpu.SideType, vx: float, vy: float, vz: float) -> None: ...
+
+class MultipleGridBuilder(LevelGridBuilder):
+ def __init__(self, *args, **kwargs) -> None: ...
+ def add_coarse_grid(self, start_x: float, start_y: float, start_z: float, end_x: float, end_y: float, end_z: float, delta: float) -> None: ...
+ @overload
+ def add_geometry(self, solid_object: Object) -> None: ...
+ @overload
+ def add_geometry(self, solid_object: Object, level: int) -> None: ...
+ @overload
+ def add_grid(self, grid_shape: Object) -> None: ...
+ @overload
+ def add_grid(self, grid_shape: Object, level_fine: int) -> None: ...
+ def build_grids(self, lbm_or_gks: pyfluids.bindings.basics.LbmOrGks, enable_thin_walls: bool) -> None: ...
+ def get_number_of_levels(self) -> int: ...
+ @staticmethod
+ def make_shared(grid_factory: GridFactory) -> MultipleGridBuilder: ...
+
+class Object:
+ def __init__(self, *args, **kwargs) -> None: ...
+
+class Sphere(Object):
+ def __init__(self, *args, **kwargs) -> None: ...
+ @staticmethod
+ def make_shared() -> Sphere: ...
+
+class TriangularMesh(Object):
+ def __init__(self, *args, **kwargs) -> None: ...
+ @staticmethod
+ def make() -> TriangularMesh: ...
diff --git a/pythonbindings/pyfluids-stubs/bindings/gpu/probes.pyi b/pythonbindings/pyfluids-stubs/bindings/gpu/probes.pyi
new file mode 100644
index 0000000000000000000000000000000000000000..af9c40078e6009efebda4450b5c5e23586aa1e83
--- /dev/null
+++ b/pythonbindings/pyfluids-stubs/bindings/gpu/probes.pyi
@@ -0,0 +1,85 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file probes.pyi
+! \ingroup gpu
+! \author Henry Korb
+=======================================================================================
+"""
+from typing import ClassVar, List
+
+import pyfluids.bindings.gpu
+
+class PlanarAverageProbe(Probe):
+ def __init__(self, probe_name: str, output_path: str, t_start_avg: int, t_start_tmp_avg: int, t_avg: int, t_start_out: int, t_out: int, plane_normal: str) -> None: ...
+
+class PlaneProbe(Probe):
+ def __init__(self, probe_name: str, output_path: str, t_start_avg: int, t_avg: int, t_start_out: int, t_out: int) -> None: ...
+ def set_probe_plane(self, pos_x: float, pos_y: float, pos_z: float, delta_x: float, delta_y: float, delta_z: float) -> None: ...
+
+class PointProbe(Probe):
+ def __init__(self, probe_name: str, output_path: str, t_start_avg: int, t_avg: int, t_start_out: int, t_out: int, output_timeseries: bool) -> None: ...
+ def add_probe_points_from_list(self, point_coords_x: List[float], point_coords_y: List[float], point_coords_z: List[float]) -> None: ...
+ def add_probe_points_from_x_normal_plane(self, pos_x: float, pos0_y: float, pos0_z: float, pos1_y: float, pos1_z: float, n_y: int, n_z: int) -> None: ...
+
+class Probe(pyfluids.bindings.gpu.PreCollisionInteractor):
+ def __init__(self, *args, **kwargs) -> None: ...
+ def add_all_available_statistics(self) -> None: ...
+ def add_statistic(self, variable: Statistic) -> None: ...
+ def set_file_name_to_n_out(self) -> None: ...
+
+class Statistic:
+ __members__: ClassVar[dict] = ... # read-only
+ Instantaneous: ClassVar[Statistic] = ...
+ Means: ClassVar[Statistic] = ...
+ SpatialCovariances: ClassVar[Statistic] = ...
+ SpatialFlatness: ClassVar[Statistic] = ...
+ SpatialMeans: ClassVar[Statistic] = ...
+ SpatialSkewness: ClassVar[Statistic] = ...
+ SpatioTemporalCovariances: ClassVar[Statistic] = ...
+ SpatioTemporalFlatness: ClassVar[Statistic] = ...
+ SpatioTemporalMeans: ClassVar[Statistic] = ...
+ SpatioTemporalSkewness: ClassVar[Statistic] = ...
+ Variances: ClassVar[Statistic] = ...
+ __entries: ClassVar[dict] = ...
+ def __init__(self, arg0: int) -> None: ...
+ def __eq__(self, arg0: object) -> bool: ...
+ def __getstate__(self) -> int: ...
+ def __hash__(self) -> int: ...
+ def __index__(self) -> int: ...
+ def __int__(self) -> int: ...
+ def __ne__(self, arg0: object) -> bool: ...
+ def __setstate__(self, arg0: int) -> None: ...
+ @property
+ def name(self) -> str: ...
+
+class WallModelProbe(Probe):
+ def __init__(self, probe_name: str, output_path: str, t_start_avg: int, t_start_tmp_avg: int, t_avg: int, t_start_out: int, t_out: int) -> None: ...
+ def set_evaluate_pressure_gradient(self, eval_press_grad: bool) -> None: ...
+ def set_force_output_to_stress(self, output_stress: bool) -> None: ...
diff --git a/pythonbindings/pyfluids-stubs/bindings/lbm.pyi b/pythonbindings/pyfluids-stubs/bindings/lbm.pyi
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/pythonbindings/pyfluids-stubs/bindings/logger.pyi b/pythonbindings/pyfluids-stubs/bindings/logger.pyi
new file mode 100644
index 0000000000000000000000000000000000000000..fe84eeb18f3245ef72ed023b2de9db7b9131d144
--- /dev/null
+++ b/pythonbindings/pyfluids-stubs/bindings/logger.pyi
@@ -0,0 +1,45 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file logger.pyi
+! \ingroup bindings
+! \author Henry Korb
+=======================================================================================
+"""
+class Logger:
+ @staticmethod
+ def change_log_path(path: str) -> None: ...
+ @staticmethod
+ def initialize_logger() -> None: ...
+
+def vf_log_critical(message: str) -> None: ...
+def vf_log_debug(message: str) -> None: ...
+def vf_log_info(message: str) -> None: ...
+def vf_log_trace(message: str) -> None: ...
+def vf_log_warning(message: str) -> None: ...
diff --git a/pythonbindings/pyfluids/__init__.py b/pythonbindings/pyfluids/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..f0537b758267e22a72e5030340de7b87d52f35c3
--- /dev/null
+++ b/pythonbindings/pyfluids/__init__.py
@@ -0,0 +1,54 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file __init__.py
+! \ingroup pyfluids
+! \author Henry Korb
+=======================================================================================
+"""
+try:
+ from .bindings import basics
+except ImportError:
+ print("Basics bindings not included")
+try:
+ from .bindings import logger
+except ImportError:
+ print("Logger bindings not included")
+try:
+ from .bindings import lbm
+except ImportError:
+ print("LBM bindings not included")
+try:
+ from .bindings import gpu
+except ImportError:
+ print("GPU bindings not included")
+try:
+ from .bindings import cpu
+except ImportError:
+ print("CPU bindings not included")
\ No newline at end of file
diff --git a/pythonbindings/pyfluids/py.typed b/pythonbindings/pyfluids/py.typed
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/pythonbindings/pymuparser/__init__.py b/pythonbindings/pymuparser/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..398069bcba03b3fe710d9d9a6398e9c530b19ee9
--- /dev/null
+++ b/pythonbindings/pymuparser/__init__.py
@@ -0,0 +1,38 @@
+r"""
+=======================================================================================
+ ____ ____ __ ______ __________ __ __ __ __
+ \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+ \ \ | | | | | |_) | | | | | | | / \ | |
+ \ \ | | | | | _ / | | | | | | / /\ \ | |
+ \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+ \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+ \ \ | | ________________________________________________________________
+ \ \ | | | ______________________________________________________________|
+ \ \| | | | __ __ __ __ ______ _______
+ \ | | |_____ | | | | | | | | | _ \ / _____)
+ \ | | _____| | | | | | | | | | | \ \ \_______
+ \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+ \ _____| |__| |________| \_______/ |__| |______/ (_______/
+
+ This file is part of VirtualFluids. VirtualFluids is free software: you can
+ redistribute it and/or modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation, either version 3 of
+ the License, or (at your option) any later version.
+
+ VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+
+! \file __init__.pyi
+! \ingroup pymuparser
+! \author Henry Korb
+=======================================================================================
+"""
+try:
+ from .bindings import Parser
+except ImportError as e:
+ raise ImportError("Pymuparser bindings were not built. Only included if VirtualFluids is built with VF_BUILD_CPU=ON.")
\ No newline at end of file
diff --git a/pythonbindings/src/VirtualFluids.cpp b/pythonbindings/src/VirtualFluids.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..20e5012e0af325440e502c704d6f372100306ab1
--- /dev/null
+++ b/pythonbindings/src/VirtualFluids.cpp
@@ -0,0 +1,63 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file VirtualFluids.cpp
+//! \ingroup src
+//! \author Henry Korb
+//=======================================================================================
+#include <pybind11/pybind11.h>
+#include "basics/basics.cpp"
+#include "lbm/lbm.cpp"
+#include "logger/logger.cpp"
+
+#ifdef VF_GPU_PYTHONBINDINGS
+#include "gpu/gpu.cpp"
+#endif
+#ifdef VF_CPU_PYTHONBINDINGS
+#include "cpu/cpu.cpp"
+#endif
+
+
+namespace py_bindings
+{
+ namespace py = pybind11;
+
+ PYBIND11_MODULE(bindings, m)
+ {
+ py::add_ostream_redirect(m, "ostream_redirect");
+ basics::makeModule(m);
+ lbm::makeModule(m);
+ logging::makeModule(m);
+#ifdef VF_GPU_PYTHONBINDINGS
+ gpu::makeModule(m);
+#endif
+#ifdef VF_CPU_PYTHONBINDINGS
+ cpu::makeModule(m);
+#endif
+ }
+}
\ No newline at end of file
diff --git a/pythonbindings/src/VirtualFluidsModulesCPU.cpp b/pythonbindings/src/VirtualFluidsModulesCPU.cpp
deleted file mode 100644
index 2fba3da494f568f7d0d0a117a579a45c9c1b9245..0000000000000000000000000000000000000000
--- a/pythonbindings/src/VirtualFluidsModulesCPU.cpp
+++ /dev/null
@@ -1,12 +0,0 @@
-#include <pybind11/pybind11.h>
-#include "cpu/cpu.cpp"
-
-namespace py_bindings
-{
- namespace py = pybind11;
-
- PYBIND11_MODULE(pyfluids, m)
- {
- cpu::makeModule(m);
- }
-}
\ No newline at end of file
diff --git a/pythonbindings/src/VirtualFluidsModulesGPU.cpp b/pythonbindings/src/VirtualFluidsModulesGPU.cpp
deleted file mode 100644
index b96971caf381faada76ee676cf60469492d055c2..0000000000000000000000000000000000000000
--- a/pythonbindings/src/VirtualFluidsModulesGPU.cpp
+++ /dev/null
@@ -1,19 +0,0 @@
-#include <pybind11/pybind11.h>
-#include "basics/basics.cpp"
-#include "lbm/lbm.cpp"
-#include "gpu/gpu.cpp"
-#include "logger/logger.cpp"
-
-namespace py_bindings
-{
- namespace py = pybind11;
-
- PYBIND11_MODULE(pyfluids, m)
- {
- basics::makeModule(m);
- gpu::makeModule(m);
- lbm::makeModule(m);
- logging::makeModule(m);
- py::add_ostream_redirect(m, "ostream_redirect");
- }
-}
\ No newline at end of file
diff --git a/pythonbindings/src/basics/basics.cpp b/pythonbindings/src/basics/basics.cpp
index 381e345d78226b25ec3a77a14340d2ef1171c8c9..e67dfb05308511c8bf79d7e860299f062f317194 100644
--- a/pythonbindings/src/basics/basics.cpp
+++ b/pythonbindings/src/basics/basics.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file basics.cpp
+//! \ingroup basics
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include "submodules/logger.cpp"
#include "submodules/configuration_file.cpp"
diff --git a/pythonbindings/src/basics/submodules/configuration_file.cpp b/pythonbindings/src/basics/submodules/configuration_file.cpp
index f5a2f87135a17f5eda34a7467d95f9db6b1c21d1..7fcd48c34824b9370eeac1872c899bf980176a52 100644
--- a/pythonbindings/src/basics/submodules/configuration_file.cpp
+++ b/pythonbindings/src/basics/submodules/configuration_file.cpp
@@ -1,5 +1,37 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file configuration_file.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
-#include <basics/config/ConfigurationFile.h>
+#include "basics/config/ConfigurationFile.h"
namespace configuration
{
@@ -9,6 +41,19 @@ namespace configuration
{
py::class_<vf::basics::ConfigurationFile>(parentModule, "ConfigurationFile")
.def(py::init<>())
- .def("load", &vf::basics::ConfigurationFile::load);
+ .def("load", &vf::basics::ConfigurationFile::load, py::arg("file"))
+ .def("contains", &vf::basics::ConfigurationFile::contains, py::arg("key"))
+ .def("get_int_value" , static_cast<int (vf::basics::ConfigurationFile::*)(const std::string&) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"))
+ .def("get_int_value" , static_cast<int (vf::basics::ConfigurationFile::*)(const std::string&, int ) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"), py::arg("default_value"))
+ .def("get_uint_value" , static_cast<uint (vf::basics::ConfigurationFile::*)(const std::string&) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"))
+ .def("get_uint_value" , static_cast<uint (vf::basics::ConfigurationFile::*)(const std::string&, uint ) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"), py::arg("default_value"))
+ .def("get_float_value" , static_cast<float (vf::basics::ConfigurationFile::*)(const std::string&) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"))
+ .def("get_float_value" , static_cast<float (vf::basics::ConfigurationFile::*)(const std::string&, float ) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"), py::arg("default_value"))
+ .def("get_double_value", static_cast<double (vf::basics::ConfigurationFile::*)(const std::string&) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"))
+ .def("get_double_value", static_cast<double (vf::basics::ConfigurationFile::*)(const std::string&, double ) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"), py::arg("default_value"))
+ .def("get_bool_value" , static_cast<bool (vf::basics::ConfigurationFile::*)(const std::string&) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"))
+ .def("get_bool_value" , static_cast<bool (vf::basics::ConfigurationFile::*)(const std::string&, bool ) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"), py::arg("default_value"))
+ .def("get_string_value", static_cast<std::string (vf::basics::ConfigurationFile::*)(const std::string&) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"))
+ .def("get_string_value", static_cast<std::string (vf::basics::ConfigurationFile::*)(const std::string&, std::string) const>(&vf::basics::ConfigurationFile::getValue), py::arg("key"), py::arg("default_value"));
}
}
\ No newline at end of file
diff --git a/pythonbindings/src/basics/submodules/lbm_or_gks.cpp b/pythonbindings/src/basics/submodules/lbm_or_gks.cpp
index ed1deeca62fc57b7f44499b306e9f99b7f990604..d20cf2d1f631f6d36a80c36f1fb6c9c59d192090 100644
--- a/pythonbindings/src/basics/submodules/lbm_or_gks.cpp
+++ b/pythonbindings/src/basics/submodules/lbm_or_gks.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file lbm_or_gks.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include "basics/Core/LbmOrGks.h"
diff --git a/pythonbindings/src/basics/submodules/logger.cpp b/pythonbindings/src/basics/submodules/logger.cpp
index d46648e349b44243581e083f3561e8a13648f3b2..fa7e00e4dca06581b7a14d2bcf2628ed6af60001 100644
--- a/pythonbindings/src/basics/submodules/logger.cpp
+++ b/pythonbindings/src/basics/submodules/logger.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file logger.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <pybind11/iostream.h>
#include <basics/Core/Logger/Logger.h>
@@ -12,12 +44,12 @@ namespace logger
py::module loggerModule = parentModule.def_submodule("logger");
py::class_<logging::Logger>(loggerModule, "Logger")
- .def("add_stdout", [](){
+ .def_static("add_stdout", [](){
logging::Logger::addStream(&std::cout);
})
- .def("set_debug_level", &logging::Logger::setDebugLevel)
- .def("time_stamp", &logging::Logger::timeStamp)
- .def("enable_printed_rank_numbers", &logging::Logger::enablePrintedRankNumbers);
+ .def_static("set_debug_level", &logging::Logger::setDebugLevel)
+ .def_static("time_stamp", &logging::Logger::timeStamp, py::arg("time_stamp"))
+ .def_static("enable_printed_rank_numbers", &logging::Logger::enablePrintedRankNumbers, py::arg("print"));
loggerModule.attr("log") = logging::out;
py::enum_<logging::Logger::Level>(loggerModule, "Level")
diff --git a/pythonbindings/src/cpu/cpu.cpp b/pythonbindings/src/cpu/cpu.cpp
index 554de53b47446366693aed31d534f6145ebea8ba..75143d913596c74a26f25ce64f1e6d214a442e34 100644
--- a/pythonbindings/src/cpu/cpu.cpp
+++ b/pythonbindings/src/cpu/cpu.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file cpu.cpp
+//! \ingroup cpu
+//! \author Sven Marcus, Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include "submodules/boundaryconditions.cpp"
#include "submodules/simulationconfig.cpp"
diff --git a/pythonbindings/src/cpu/submodules/boundaryconditions.cpp b/pythonbindings/src/cpu/submodules/boundaryconditions.cpp
index 3bff7bc069ca20fe1c0cf3d1847b9714e0381505..ac9ec8605dec51e8374c850b1c1b58314674c426 100644
--- a/pythonbindings/src/cpu/submodules/boundaryconditions.cpp
+++ b/pythonbindings/src/cpu/submodules/boundaryconditions.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file boundaryconditions.cpp
+//! \ingroup submodules
+//! \author Sven Marcus, Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <pybind11/stl.h>
#include <BoundaryConditions/DensityBCAdapter.h>
diff --git a/pythonbindings/src/cpu/submodules/geometry.cpp b/pythonbindings/src/cpu/submodules/geometry.cpp
index b7ff4dd761258d41687589d2dd89c3479093753e..4c4c47b002b9c7451a8d788ba82c4a19b78ca96f 100644
--- a/pythonbindings/src/cpu/submodules/geometry.cpp
+++ b/pythonbindings/src/cpu/submodules/geometry.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file geometry.cpp
+//! \ingroup submodules
+//! \author Sven Marcus, Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <geometry3d/GbPoint3D.h>
#include <geometry3d/GbObject3D.h>
diff --git a/pythonbindings/src/cpu/submodules/kernel.cpp b/pythonbindings/src/cpu/submodules/kernel.cpp
index fb291790632cc2041410f60a14fca8d966283343..b00d86579540a299e4bf3ed47bc09d4386f420a2 100644
--- a/pythonbindings/src/cpu/submodules/kernel.cpp
+++ b/pythonbindings/src/cpu/submodules/kernel.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file kernel.cpp
+//! \ingroup submodules
+//! \author Sven Marcus, Henry Korb
+//=======================================================================================
#include <memory>
#include <pybind11/pybind11.h>
#include <simulationconfig/KernelFactory.h>
diff --git a/pythonbindings/src/cpu/submodules/simulationconfig.cpp b/pythonbindings/src/cpu/submodules/simulationconfig.cpp
index 60af4e36af4dca67e9262dd9f5ee1f46d5b7bb58..09d91f44e85f03c6150c56ce5762e7629212fba0 100644
--- a/pythonbindings/src/cpu/submodules/simulationconfig.cpp
+++ b/pythonbindings/src/cpu/submodules/simulationconfig.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file simulationconfig.cpp
+//! \ingroup submodules
+//! \author Sven Marcus, Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <simulationconfig/Simulation.h>
diff --git a/pythonbindings/src/cpu/submodules/simulationparameters.cpp b/pythonbindings/src/cpu/submodules/simulationparameters.cpp
index acc272f2ee412cfbafd9007b4b18610cfd0a1e9b..b33d20f9e5d335a0ed381faf8786d88cc7642738 100644
--- a/pythonbindings/src/cpu/submodules/simulationparameters.cpp
+++ b/pythonbindings/src/cpu/submodules/simulationparameters.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file simulationparameters.cpp
+//! \ingroup submodules
+//! \author Sven Marcus, Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <pybind11/stl.h>
#include <complex>
diff --git a/pythonbindings/src/cpu/submodules/writer.cpp b/pythonbindings/src/cpu/submodules/writer.cpp
index d5ec527a27caf63d9a3066c51e1f675b307fe0b2..f1cfd8934c2da84266a93d5bcd91eb26f5f69d3f 100644
--- a/pythonbindings/src/cpu/submodules/writer.cpp
+++ b/pythonbindings/src/cpu/submodules/writer.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file writer.cpp
+//! \ingroup submodules
+//! \author Sven Marcus, Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <simulationconfig/WriterConfiguration.h>
diff --git a/pythonbindings/src/gpu/gpu.cpp b/pythonbindings/src/gpu/gpu.cpp
index dc110cd5e19a9aad4937f9c2133ddf74c0ddf9bf..9eb160ae7765f16a6437e343cb878bb4b80877bf 100644
--- a/pythonbindings/src/gpu/gpu.cpp
+++ b/pythonbindings/src/gpu/gpu.cpp
@@ -1,14 +1,50 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file gpu.cpp
+//! \ingroup gpu
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
-#include "submodules/actuator_line.cpp"
#include "submodules/pre_collision_interactor.cpp"
#include "submodules/simulation.cpp"
#include "submodules/parameter.cpp"
#include "submodules/boundary_conditions.cpp"
#include "submodules/communicator.cpp"
#include "submodules/cuda_memory_manager.cpp"
+#include "submodules/probes.cpp"
+#include "submodules/precursor_writer.cpp"
#include "submodules/grid_provider.cpp"
#include "submodules/grid_generator.cpp"
-#include "submodules/probes.cpp"
+#include "submodules/turbulence_models.cpp"
+#include "submodules/transient_bc_setter.cpp"
+#include "submodules/actuator_farm.cpp"
+#include "submodules/grid_scaling_factory.cpp"
namespace gpu
{
@@ -20,13 +56,17 @@ namespace gpu
simulation::makeModule(gpuModule);
parameter::makeModule(gpuModule);
pre_collision_interactor::makeModule(gpuModule);
- actuator_line::makeModule(gpuModule);
+ actuator_farm::makeModule(gpuModule);
boundary_conditions::makeModule(gpuModule);
+ transient_bc_setter::makeModule(gpuModule);
communicator::makeModule(gpuModule);
cuda_memory_manager::makeModule(gpuModule);
- grid_provider::makeModule(gpuModule);
probes::makeModule(gpuModule);
+ precursor_writer::makeModule(gpuModule);
grid_generator::makeModule(gpuModule);
+ grid_provider::makeModule(gpuModule);
+ turbulence_model::makeModule(gpuModule);
+ grid_scaling_factory::makeModule(gpuModule);
return gpuModule;
}
}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/actuator_farm.cpp b/pythonbindings/src/gpu/submodules/actuator_farm.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..a930616db3e0d0713bdf57157387d75d171603de
--- /dev/null
+++ b/pythonbindings/src/gpu/submodules/actuator_farm.cpp
@@ -0,0 +1,171 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file actuator_farm.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
+#include <pybind11/pybind11.h>
+#include <pybind11/numpy.h>
+#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorFarm.h>
+#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h>
+class PyActuatorFarm : public ActuatorFarm
+{
+public:
+ using ActuatorFarm::ActuatorFarm; // Inherit constructors
+ void calcBladeForces() override
+ {
+ PYBIND11_OVERRIDE_NAME(void, ActuatorFarm, "calc_blade_forces", calcBladeForces);
+ }
+};
+namespace actuator_farm
+{
+ namespace py = pybind11;
+
+ void makeModule(py::module_ &parentModule)
+ {
+ using arr = py::array_t<float, py::array::c_style>;
+
+ py::class_<ActuatorFarm, PreCollisionInteractor, PyActuatorFarm, std::shared_ptr<ActuatorFarm>>(parentModule, "ActuatorFarm", py::dynamic_attr())
+ .def(py::init< const uint,
+ const real,
+ const uint,
+ const real,
+ int,
+ const real,
+ const real,
+ const bool>(),
+ py::arg("number_of_blades_per_turbine"),
+ py::arg("density"),
+ py::arg("number_of_nodes_per_blade"),
+ py::arg("epsilon"),
+ py::arg("level"),
+ py::arg("delta_t"),
+ py::arg("delta_x"),
+ py::arg("use_host_arrays"))
+ .def_property_readonly("number_of_turbines", &ActuatorFarm::getNumberOfTurbines)
+ .def_property_readonly("number_of_nodes_per_blade", &ActuatorFarm::getNumberOfNodesPerBlade)
+ .def_property_readonly("number_of_blades_per_turbine", &ActuatorFarm::getNumberOfBladesPerTurbine)
+ .def_property_readonly("number_of_nodes", &ActuatorFarm::getNumberOfNodes)
+ .def_property_readonly("number_of_indices", &ActuatorFarm::getNumberOfIndices)
+ .def_property_readonly("density", &ActuatorFarm::getDensity)
+ .def_property_readonly("delta_t", &ActuatorFarm::getDeltaT)
+ .def_property_readonly("delta_x", &ActuatorFarm::getDeltaX)
+
+ .def("add_turbine", &ActuatorFarm::addTurbine, py::arg("posX"), py::arg("posY"), py::arg("posZ"), py::arg("diameter"), py::arg("omega"), py::arg("azimuth"), py::arg("yaw"), py::arg("bladeRadii"))
+
+ .def("get_turbine_pos", [](ActuatorFarm& al, uint turbine){ real position[3] = {al.getTurbinePosX(turbine), al.getTurbinePosY(turbine), al.getTurbinePosZ(turbine)}; return arr(3, position); }, py::arg("turbine"))
+ .def("get_turbine_azimuth", &ActuatorFarm::getTurbineAzimuth, py::arg("turbine"))
+ .def("get_turbine_yaw", &ActuatorFarm::getTurbineYaw, py::arg("turbine"))
+ .def("get_turbine_omega", &ActuatorFarm::getTurbineOmega, py::arg("turbine"))
+ .def("get_all_azimuths", [](ActuatorFarm& al){ return arr(al.getNumberOfTurbines(), al.getAllAzimuths()); } )
+ .def("get_all_yaws", [](ActuatorFarm& al){ return arr(al.getNumberOfTurbines(), al.getAllYaws()); } )
+ .def("get_all_omegas", [](ActuatorFarm& al){ return arr(al.getNumberOfTurbines(), al.getAllOmegas()); } )
+ .def("get_all_turbine_pos_x", [](ActuatorFarm& al){ return arr(al.getNumberOfTurbines(), al.getAllTurbinePosX()); } )
+ .def("get_all_turbine_pos_y", [](ActuatorFarm& al){ return arr(al.getNumberOfTurbines(), al.getAllTurbinePosY()); } )
+ .def("get_all_turbine_pos_z", [](ActuatorFarm& al){ return arr(al.getNumberOfTurbines(), al.getAllTurbinePosZ()); } )
+
+ .def("get_all_blade_radii", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfNodesPerBlade()}, al.getAllBladeRadii()); } )
+ .def("get_all_blade_coords_x", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeCoordsX()); } )
+ .def("get_all_blade_coords_y", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeCoordsY()); } )
+ .def("get_all_blade_coords_z", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeCoordsZ()); } )
+ .def("get_all_blade_velocities_x", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeVelocitiesX()); } )
+ .def("get_all_blade_velocities_y", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeVelocitiesY()); } )
+ .def("get_all_blade_velocities_z", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeVelocitiesZ()); } )
+ .def("get_all_blade_forces_x", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeForcesX()); } )
+ .def("get_all_blade_forces_y", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeForcesY()); } )
+ .def("get_all_blade_forces_z", [](ActuatorFarm& al){ return arr({al.getNumberOfTurbines(), al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getAllBladeForcesZ()); } )
+
+ .def("get_turbine_blade_radii", [](ActuatorFarm& al, uint turbine){ return arr(al.getNumberOfNodesPerBlade(), al.getTurbineBladeRadiiDevice(turbine)); } , py::arg("turbine"))
+ .def("get_turbine_blade_coords_x", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeCoordsXDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_coords_y", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeCoordsYDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_coords_z", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeCoordsZDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_velocities_x", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeVelocitiesXDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_velocities_y", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeVelocitiesYDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_velocities_z", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeVelocitiesZDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_forces_x", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeForcesXDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_forces_y", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeForcesYDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_forces_z", [](ActuatorFarm& al, uint turbine){ return arr({al.getNumberOfBladesPerTurbine(), al.getNumberOfNodesPerBlade()}, al.getTurbineBladeForcesZDevice(turbine)); }, py::arg("turbine") )
+
+ .def("get_all_blade_radii_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t>(al.getAllBladeRadiiDevice()); } )
+ .def("get_all_blade_coords_x_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeCoordsXDevice()); } )
+ .def("get_all_blade_coords_y_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeCoordsYDevice()); } )
+ .def("get_all_blade_coords_z_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeCoordsZDevice()); } )
+ .def("get_all_blade_velocities_x_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeVelocitiesXDevice()); } )
+ .def("get_all_blade_velocities_y_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeVelocitiesYDevice()); } )
+ .def("get_all_blade_velocities_z_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeVelocitiesZDevice()); } )
+ .def("get_all_blade_forces_x_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeForcesXDevice()); } )
+ .def("get_all_blade_forces_y_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeForcesYDevice()); } )
+ .def("get_all_blade_forces_z_device", [](ActuatorFarm& al) -> intptr_t { return reinterpret_cast<intptr_t> (al.getAllBladeForcesZDevice()); } )
+
+ .def("get_turbine_blade_radii_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeRadiiDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_coords_x_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeCoordsXDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_coords_y_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeCoordsYDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_coords_z_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeCoordsZDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_velocities_x_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeVelocitiesXDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_velocities_y_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeVelocitiesYDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_velocities_z_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeVelocitiesZDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_forces_x_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeForcesXDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_forces_y_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeForcesYDevice(turbine)); }, py::arg("turbine") )
+ .def("get_turbine_blade_forces_z_device", [](ActuatorFarm& al, uint turbine) -> intptr_t { return reinterpret_cast<intptr_t>(al.getTurbineBladeForcesZDevice(turbine)); }, py::arg("turbine") )
+
+ .def("set_all_azimuths", [](ActuatorFarm& al, arr azimuths){ al.setAllAzimuths(static_cast<float *>(azimuths.request().ptr)); }, py::arg("azimuths"))
+ .def("set_all_yaws", [](ActuatorFarm& al, arr yaws){ al.setAllYaws(static_cast<float *>(yaws.request().ptr)); }, py::arg("yaws"))
+ .def("set_all_omegas", [](ActuatorFarm& al, arr omegas){ al.setAllOmegas(static_cast<float *>(omegas.request().ptr)); }, py::arg("omegas"))
+
+ .def("set_turbine_azimuth", &ActuatorFarm::setTurbineAzimuth, py::arg("turbine"), py::arg("azimuth"))
+ .def("set_turbine_yaw", &ActuatorFarm::setTurbineYaw, py::arg("turbine"), py::arg("yaw"))
+ .def("set_turbine_omega", &ActuatorFarm::setTurbineOmega, py::arg("turbine"), py::arg("omega"))
+
+ .def("set_all_blade_coords", [](ActuatorFarm& al, arr coordsX, arr coordsY, arr coordsZ)
+ {
+ al.setAllBladeCoords(static_cast<float *>(coordsX.request().ptr), static_cast<float *>(coordsY.request().ptr), static_cast<float *>(coordsZ.request().ptr));
+ }, py::arg("blade_coords_x"), py::arg("blade_coords_y"), py::arg("blade_coords_z") )
+ .def("set_all_blade_velocities", [](ActuatorFarm& al, arr velocitiesX, arr velocitiesY, arr velocitiesZ)
+ {
+ al.setAllBladeVelocities(static_cast<float *>(velocitiesX.request().ptr), static_cast<float *>(velocitiesY.request().ptr), static_cast<float *>(velocitiesZ.request().ptr));
+ }, py::arg("blade_velocities_x"), py::arg("blade_velocities_y"), py::arg("blade_velocities_z") )
+ .def("set_all_blade_forces", [](ActuatorFarm& al, arr forcesX, arr forcesY, arr forcesZ)
+ {
+ al.setAllBladeForces(static_cast<float *>(forcesX.request().ptr), static_cast<float *>(forcesY.request().ptr), static_cast<float *>(forcesZ.request().ptr));
+ }, py::arg("blade_forces_x"), py::arg("blade_forces_y"), py::arg("blade_forces_z") )
+ .def("set_turbine_blade_coords", [](ActuatorFarm& al, uint turbine, arr coordsX, arr coordsY, arr coordsZ)
+ {
+ al.setTurbineBladeCoords(turbine, static_cast<float *>(coordsX.request().ptr), static_cast<float *>(coordsY.request().ptr), static_cast<float *>(coordsZ.request().ptr));
+ }, py::arg("turbine"), py::arg("blade_coords_x"), py::arg("blade_coords_y"), py::arg("blade_coords_z") )
+ .def("set_turbine_blade_velocities", [](ActuatorFarm& al, uint turbine, arr velocitiesX, arr velocitiesY, arr velocitiesZ)
+ {
+ al.setTurbineBladeVelocities(turbine, static_cast<float *>(velocitiesX.request().ptr), static_cast<float *>(velocitiesY.request().ptr), static_cast<float *>(velocitiesZ.request().ptr));
+ }, py::arg("turbine"), py::arg("blade_velocities_x"), py::arg("blade_velocities_y"), py::arg("blade_velocities_z") )
+ .def("set_turbine_blade_forces", [](ActuatorFarm& al, uint turbine, arr forcesX, arr forcesY, arr forcesZ)
+ {
+ al.setTurbineBladeForces(turbine, static_cast<float *>(forcesX.request().ptr), static_cast<float *>(forcesY.request().ptr), static_cast<float *>(forcesZ.request().ptr));
+ }, py::arg("turbine"), py::arg("blade_forces_x"), py::arg("blade_forces_y"), py::arg("blade_forces_z") )
+ .def("calc_blade_forces", &ActuatorFarm::calcBladeForces);
+ }
+}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/actuator_line.cpp b/pythonbindings/src/gpu/submodules/actuator_line.cpp
deleted file mode 100644
index 3207fadbc37df38e53e00adcb9a86f0b8e82ba98..0000000000000000000000000000000000000000
--- a/pythonbindings/src/gpu/submodules/actuator_line.cpp
+++ /dev/null
@@ -1,72 +0,0 @@
-#include <pybind11/pybind11.h>
-#include <pybind11/stl.h>
-#include <pybind11/numpy.h>
-#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.h>
-#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h>
-class PyActuatorLine : public ActuatorLine
-{
-public:
- using ActuatorLine::ActuatorLine; // Inherit constructors
- void calcBladeForces() override
- {
- PYBIND11_OVERRIDE_NAME(void, ActuatorLine, "calc_blade_forces", calcBladeForces,);
- }
-};
-namespace actuator_line
-{
- namespace py = pybind11;
-
- void makeModule(py::module_ &parentModule)
- {
- using arr = py::array_t<float, py::array::c_style>;
-
- py::class_<ActuatorLine, PreCollisionInteractor, PyActuatorLine, std::shared_ptr<ActuatorLine>>(parentModule, "ActuatorLine", py::dynamic_attr())
- .def(py::init< const uint,
- const real,
- const uint,
- const real,
- real, real, real,
- const real,
- int,
- const real,
- const real>(),
- "n_blades",
- "density",
- "n_blade_nodes",
- "epsilon",
- "turbine_pos_x", "turbine_pos_y", "turbine_pos_z",
- "diameter",
- "level",
- "delta_t",
- "delta_x")
- .def_property("omega", &ActuatorLine::getOmega, &ActuatorLine::setOmega)
- .def_property("azimuth", &ActuatorLine::getAzimuth, &ActuatorLine::setAzimuth)
- .def_property("yaw", &ActuatorLine::getYaw, &ActuatorLine::setYaw)
- .def_property_readonly("n_blades", &ActuatorLine::getNBlades)
- .def_property_readonly("n_blade_nodes", &ActuatorLine::getNBladeNodes)
- .def_property_readonly("n_nodes", &ActuatorLine::getNNodes)
- .def_property_readonly("n_indices", &ActuatorLine::getNIndices)
- .def_property_readonly("density", &ActuatorLine::getDensity)
- .def_property_readonly("position_x", &ActuatorLine::getPositionX)
- .def_property_readonly("position_y", &ActuatorLine::getPositionY)
- .def_property_readonly("position_z", &ActuatorLine::getPositionZ)
- .def_property_readonly("position", [](ActuatorLine& al){ real position[3] = {al.getPositionX(), al.getPositionY(), al.getPositionZ()}; return arr(3, position); } )
- .def("get_radii", [](ActuatorLine& al){ return arr(al.getNBladeNodes(), al.getBladeRadii()); } )
- .def("get_blade_coords_x", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeCoordsX()); } )
- .def("get_blade_coords_y", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeCoordsY()); } )
- .def("get_blade_coords_z", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeCoordsZ()); } )
- .def("get_blade_velocities_x", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeVelocitiesX()); } )
- .def("get_blade_velocities_y", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeVelocitiesY()); } )
- .def("get_blade_velocities_z", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeVelocitiesZ()); } )
- .def("get_blade_forces_x", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeForcesX()); } )
- .def("get_blade_forces_y", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeForcesY()); } )
- .def("get_blade_forces_z", [](ActuatorLine& al){ return arr({al.getNBlades(), al.getNBladeNodes()}, al.getBladeForcesZ()); } )
- .def("set_blade_coords", [](ActuatorLine& al, arr coordsX, arr coordsY, arr coordsZ){
- al.setBladeCoords(static_cast<float *>(coordsX.request().ptr), static_cast<float *>(coordsY.request().ptr), static_cast<float *>(coordsZ.request().ptr)); } )
- .def("set_blade_velocities", [](ActuatorLine& al, arr velocitiesX, arr velocitiesY, arr velocitiesZ){
- al.setBladeVelocities(static_cast<float *>(velocitiesX.request().ptr), static_cast<float *>(velocitiesY.request().ptr), static_cast<float *>(velocitiesZ.request().ptr)); } )
- .def("set_blade_forces", [](ActuatorLine& al, arr forcesX, arr forcesY, arr forcesZ){
- al.setBladeForces(static_cast<float *>(forcesX.request().ptr), static_cast<float *>(forcesY.request().ptr), static_cast<float *>(forcesZ.request().ptr)); } )
- .def("calc_blade_forces", &ActuatorLine::calcBladeForces);
- }
-}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/boundary_conditions.cpp b/pythonbindings/src/gpu/submodules/boundary_conditions.cpp
index 8f941a8705c225275d25291205ebdaeef8de5c9e..865817bb16f7b164c40bdc066645fb2e1f1c842e 100644
--- a/pythonbindings/src/gpu/submodules/boundary_conditions.cpp
+++ b/pythonbindings/src/gpu/submodules/boundary_conditions.cpp
@@ -1,5 +1,38 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file boindary_conditions.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <gpu/GridGenerator/grid/BoundaryConditions/Side.h>
+#include "gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.h"
namespace boundary_conditions
{
@@ -14,7 +47,59 @@ namespace boundary_conditions
.value("PY", SideType::PY)
.value("MZ", SideType::MZ)
.value("PZ", SideType::PZ)
- .value("GEOMETRY", SideType::GEOMETRY)
- .export_values();
+ .value("GEOMETRY", SideType::GEOMETRY);
+
+ py::class_<BoundaryConditionFactory>(parentModule, "BoundaryConditionFactory")
+ .def(py::init<>())
+ .def("set_velocity_boundary_condition", &BoundaryConditionFactory::setVelocityBoundaryCondition, py::arg("boundary_condition_type"))
+ .def("set_no_slip_boundary_condition", &BoundaryConditionFactory::setNoSlipBoundaryCondition, py::arg("boundary_condition_type"))
+ .def("set_slip_boundary_condition", &BoundaryConditionFactory::setSlipBoundaryCondition, py::arg("boundary_condition_type"))
+ .def("set_pressure_boundary_condition", &BoundaryConditionFactory::setPressureBoundaryCondition, py::arg("boundary_condition_type"))
+ .def("set_stress_boundary_condition", &BoundaryConditionFactory::setStressBoundaryCondition, py::arg("boundary_condition_type"))
+ .def("set_precursor_boundary_condition", &BoundaryConditionFactory::setPrecursorBoundaryCondition, py::arg("boundary_condition_type"))
+ .def("set_geometry_boundary_condition", &BoundaryConditionFactory::setGeometryBoundaryCondition, py::arg("boundary_condition_type"));
+
+ py::enum_<BoundaryConditionFactory::VelocityBC>(parentModule, "VelocityBC")
+ .value("VelocitySimpleBounceBackCompressible", BoundaryConditionFactory::VelocityBC::VelocitySimpleBounceBackCompressible)
+ .value("VelocityIncompressible", BoundaryConditionFactory::VelocityBC::VelocityIncompressible)
+ .value("VelocityCompressible", BoundaryConditionFactory::VelocityBC::VelocityCompressible)
+ .value("VelocityAndPressureCompressible", BoundaryConditionFactory::VelocityBC::VelocityAndPressureCompressible)
+ .value("NotSpecified", BoundaryConditionFactory::VelocityBC::NotSpecified);
+
+
+ py::enum_<BoundaryConditionFactory::NoSlipBC>(parentModule, "NoSlipBC")
+ .value("NoSlipImplicitBounceBack", BoundaryConditionFactory::NoSlipBC::NoSlipImplicitBounceBack)
+ .value("NoSlipBounceBack", BoundaryConditionFactory::NoSlipBC::NoSlipBounceBack)
+ .value("NoSlipIncompressible", BoundaryConditionFactory::NoSlipBC::NoSlipIncompressible)
+ .value("NoSlipCompressible", BoundaryConditionFactory::NoSlipBC::NoSlipCompressible)
+ .value("NoSlip3rdMomentsCompressible", BoundaryConditionFactory::NoSlipBC::NoSlip3rdMomentsCompressible);
+
+ py::enum_<BoundaryConditionFactory::SlipBC>(parentModule, "SlipBC")
+ .value("SlipIncompressible", BoundaryConditionFactory::SlipBC::SlipIncompressible)
+ .value("SlipCompressible", BoundaryConditionFactory::SlipBC::SlipCompressible)
+ .value("SlipBounceBack", BoundaryConditionFactory::SlipBC::SlipBounceBack)
+ .value("SlipCompressibleTurbulentViscosity", BoundaryConditionFactory::SlipBC::SlipCompressibleTurbulentViscosity)
+ .value("SlipPressureCompressibleTurbulentViscosity", BoundaryConditionFactory::SlipBC::SlipPressureCompressibleTurbulentViscosity)
+ .value("NotSpecified", BoundaryConditionFactory::SlipBC::NotSpecified);
+
+ py::enum_<BoundaryConditionFactory::PressureBC>(parentModule, "PressureBC")
+ .value("PressureEquilibrium", BoundaryConditionFactory::PressureBC::PressureEquilibrium)
+ .value("PressureEquilibrium2", BoundaryConditionFactory::PressureBC::PressureEquilibrium2)
+ .value("PressureNonEquilibriumIncompressible", BoundaryConditionFactory::PressureBC::PressureNonEquilibriumIncompressible)
+ .value("PressureNonEquilibriumCompressible", BoundaryConditionFactory::PressureBC::PressureNonEquilibriumCompressible)
+ .value("OutflowNonReflective", BoundaryConditionFactory::PressureBC::OutflowNonReflective)
+ .value("OutflowNonReflectivePressureCorrection", BoundaryConditionFactory::PressureBC::OutflowNonReflectivePressureCorrection)
+ .value("NotSpecified", BoundaryConditionFactory::PressureBC::NotSpecified);
+
+ py::enum_<BoundaryConditionFactory::StressBC>(parentModule, "StressBC")
+ .value("StressCompressible", BoundaryConditionFactory::StressBC::StressCompressible)
+ .value("StressBounceBack", BoundaryConditionFactory::StressBC::StressBounceBack)
+ .value("StressPressureBounceBack", BoundaryConditionFactory::StressBC::StressPressureBounceBack)
+ .value("NotSpecified", BoundaryConditionFactory::StressBC::NotSpecified);
+
+ py::enum_<BoundaryConditionFactory::PrecursorBC>(parentModule, "PrecursorBC")
+ .value("VelocityPrecursor", BoundaryConditionFactory::PrecursorBC::VelocityPrecursor)
+ .value("DistributionsPrecursor", BoundaryConditionFactory::PrecursorBC::DistributionsPrecursor)
+ .value("NotSpecified", BoundaryConditionFactory::PrecursorBC::NotSpecified);
}
}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/communicator.cpp b/pythonbindings/src/gpu/submodules/communicator.cpp
index edb36e2c2f774903590a16a0b406c721662827b1..26a57061933fbdbfe3447ec89eeb07116a9b974b 100644
--- a/pythonbindings/src/gpu/submodules/communicator.cpp
+++ b/pythonbindings/src/gpu/submodules/communicator.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file communicator.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <gpu/VirtualFluids_GPU/Communication/Communicator.h>
@@ -8,7 +40,7 @@ namespace communicator
void makeModule(py::module_ &parentModule)
{
py::class_<vf::gpu::Communicator, std::unique_ptr<vf::gpu::Communicator, py::nodelete>>(parentModule, "Communicator")
- .def("get_instance", &vf::gpu::Communicator::getInstance, py::return_value_policy::reference)
+ .def_static("get_instance", &vf::gpu::Communicator::getInstance, py::return_value_policy::reference)
.def("get_number_of_process", &vf::gpu::Communicator::getNummberOfProcess)
.def("get_pid", &vf::gpu::Communicator::getPID);
}
diff --git a/pythonbindings/src/gpu/submodules/cuda_memory_manager.cpp b/pythonbindings/src/gpu/submodules/cuda_memory_manager.cpp
index bf27080cb3cd050343ba42b0571827ed58870cfd..bbff4832cb73f47e3d1a5a6abd78e21da2473deb 100644
--- a/pythonbindings/src/gpu/submodules/cuda_memory_manager.cpp
+++ b/pythonbindings/src/gpu/submodules/cuda_memory_manager.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file cuda_memory_manager.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.h>
#include <gpu/VirtualFluids_GPU/Parameter/Parameter.h>
@@ -10,6 +42,6 @@ namespace cuda_memory_manager
void makeModule(py::module_ &parentModule)
{
py::class_<CudaMemoryManager, std::shared_ptr<CudaMemoryManager>>(parentModule, "CudaMemoryManager")
- .def(py::init<std::shared_ptr<Parameter>>(), "parameter");
+ .def(py::init<std::shared_ptr<Parameter>>(), py::arg("parameter"));
}
}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/grid_generator.cpp b/pythonbindings/src/gpu/submodules/grid_generator.cpp
index 579c06c4e00cae9646ced8b554d71631eeb7e793..3e9fb5655e26ffa6053a205da5a3e3f0f2ecd49f 100644
--- a/pythonbindings/src/gpu/submodules/grid_generator.cpp
+++ b/pythonbindings/src/gpu/submodules/grid_generator.cpp
@@ -1,4 +1,37 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file grid_generator.cpp
+//! \ingroup submodules
+//! \author Henry Korb, Henrik Asmuth
+//=======================================================================================
#include <pybind11/pybind11.h>
+#include "gpu/GridGenerator/utilities/communication.h"
#include "gpu/GridGenerator/geometries/Object.h"
#include "gpu/GridGenerator/geometries/BoundingBox/BoundingBox.h"
#include "gpu/GridGenerator/geometries/Conglomerate/Conglomerate.h"
@@ -17,51 +50,63 @@ namespace grid_generator
{
py::module gridGeneratorModule = parentModule.def_submodule("grid_generator");
+ //TODO:
+ // py::enum_<CommunicationDirections>(gridGeneratorModule, "CommunicationDirections")
+ // .value("MX", CommunicationDirections::MX)
+ // .value("PX", CommunicationDirections::PX)
+ // .value("MY", CommunicationDirections::MY)
+ // .value("PY", CommunicationDirections::PY)
+ // .value("MZ", CommunicationDirections::MZ)
+ // .value("PZ", CommunicationDirections::PZ);
+
py::class_<GridFactory, std::shared_ptr<GridFactory>>(gridGeneratorModule, "GridFactory")
- .def("make", &GridFactory::make, py::return_value_policy::reference);
+ .def_static("make", &GridFactory::make, py::return_value_policy::reference);
- py::class_<BoundingBox>(gridGeneratorModule, "BoundingBox")
- .def(py::init<real, real, real, real, real, real>(),"min_x","max_x","min_y","max_y","min_z","max_z");
+ py::class_<BoundingBox, std::shared_ptr<BoundingBox>>(gridGeneratorModule, "BoundingBox")
+ .def(py::init<real, real, real, real, real, real>(), py::arg("min_x"), py::arg("max_x"), py::arg("min_y"), py::arg("max_y"), py::arg("min_z"), py::arg("max_z"));
py::class_<Object, std::shared_ptr<Object>>(gridGeneratorModule, "Object");
py::class_<Conglomerate, Object, std::shared_ptr<Conglomerate>>(gridGeneratorModule, "Conglomerate")
- .def("make_shared", &Conglomerate::makeShared, py::return_value_policy::reference)
- .def("add", &Conglomerate::add)
- .def("subtract", &Conglomerate::subtract);
+ .def_static("make_shared", &Conglomerate::makeShared, py::return_value_policy::reference)
+ .def("add", &Conglomerate::add, py::arg("object"))
+ .def("subtract", &Conglomerate::subtract, py::arg("object"));
py::class_<Cuboid, Object, std::shared_ptr<Cuboid>>(gridGeneratorModule, "Cuboid")
.def(py::init<const double&, const double&, const double&, const double&, const double&, const double&>(),
- "min_x1", "min_x2", "min_x3", "max_x1", "max_x2", "max_x3");
+ py::arg("min_x1"), py::arg("min_x2"), py::arg("min_x3"), py::arg("max_x1"), py::arg("max_x2"), py::arg("max_x3"));
py::class_<Sphere, Object, std::shared_ptr<Sphere>>(gridGeneratorModule, "Sphere")
- .def("make_shared", &Sphere::makeShared, py::return_value_policy::reference);
+ .def_static("make_shared", &Sphere::makeShared, py::return_value_policy::reference);
py::class_<TriangularMesh, Object, std::shared_ptr<TriangularMesh>>(gridGeneratorModule, "TriangularMesh")
- .def("make", &TriangularMesh::make, py::return_value_policy::reference);
+ .def_static("make", &TriangularMesh::make, py::return_value_policy::reference);
py::class_<GridBuilder, std::shared_ptr<GridBuilder>>(gridGeneratorModule, "GridBuilder")
- .def("get_number_of_grid_levels", &GridBuilder::getNumberOfGridLevels)
- .def("get_grid", &GridBuilder::getGrid);
+ .def("get_number_of_grid_levels", &GridBuilder::getNumberOfGridLevels);
py::class_<LevelGridBuilder, GridBuilder, std::shared_ptr<LevelGridBuilder>>(gridGeneratorModule, "LevelGridBuilder")
- .def("get_grid", py::overload_cast<int, int>(&LevelGridBuilder::getGrid))
- .def("set_slip_boundary_condition", &LevelGridBuilder::setSlipBoundaryCondition)
- .def("set_velocity_boundary_condition", &LevelGridBuilder::setVelocityBoundaryCondition)
- .def("set_pressure_boundary_condition", &LevelGridBuilder::setPressureBoundaryCondition)
- .def("set_periodic_boundary_condition", &LevelGridBuilder::setPeriodicBoundaryCondition)
- .def("set_no_slip_boundary_condition", &LevelGridBuilder::setNoSlipBoundaryCondition)
- .def("set_stress_boundary_condition", &LevelGridBuilder::setStressBoundaryCondition);
+ .def("set_slip_boundary_condition", &LevelGridBuilder::setSlipBoundaryCondition, py::arg("side_type"), py::arg("normal_x"), py::arg("normal_y"), py::arg("normal_z"))
+ .def("set_velocity_boundary_condition", &LevelGridBuilder::setVelocityBoundaryCondition, py::arg("side_type"), py::arg("vx"), py::arg("vy"), py::arg("vz"))
+ .def("set_pressure_boundary_condition", &LevelGridBuilder::setPressureBoundaryCondition, py::arg("side_type"), py::arg("rho"))
+ .def("set_periodic_boundary_condition", &LevelGridBuilder::setPeriodicBoundaryCondition, py::arg("periodic_x"), py::arg("periodic_y"), py::arg("periodic_z"))
+ .def("set_no_slip_boundary_condition", &LevelGridBuilder::setNoSlipBoundaryCondition, py::arg("side_type"))
+ .def("set_precursor_boundary_condition", &LevelGridBuilder::setPrecursorBoundaryCondition, py::arg("side_type"), py::arg("file_collection"), py::arg("n_t_read"), py::arg("velocity_x")=0.0f, py::arg("velocity_y")=0.0f, py::arg("velocity_z")=0.0f, py::arg("file_level_to_grid_level_map")=std::vector<uint>())
+ .def("set_stress_boundary_condition", &LevelGridBuilder::setStressBoundaryCondition, py::arg("side_type"), py::arg("normal_x"), py::arg("normal_y"), py::arg("normal_z"), py::arg("sampling_offset"), py::arg("z0"), py::arg("dx"));
py::class_<MultipleGridBuilder, LevelGridBuilder, std::shared_ptr<MultipleGridBuilder>>(gridGeneratorModule, "MultipleGridBuilder")
- .def("make_shared", &MultipleGridBuilder::makeShared, py::return_value_policy::reference)
- .def("add_coarse_grid", &MultipleGridBuilder::addCoarseGrid)
- .def("add_grid", py::overload_cast<Object*>(&MultipleGridBuilder::addGrid))
- .def("add_grid", py::overload_cast<Object*, uint>(&MultipleGridBuilder::addGrid))
- .def("add_geometry", py::overload_cast<Object*>(&MultipleGridBuilder::addGeometry))
- .def("add_geometry", py::overload_cast<Object*, uint>(&MultipleGridBuilder::addGeometry))
+ .def_static("make_shared", &MultipleGridBuilder::makeShared, py::return_value_policy::reference, py::arg("grid_factory"))
+ .def("add_coarse_grid", &MultipleGridBuilder::addCoarseGrid, py::arg("start_x"), py::arg("start_y"), py::arg("start_z"), py::arg("end_x"), py::arg("end_y"), py::arg("end_z"), py::arg("delta"))
+ .def("add_grid", py::overload_cast<Object*>(&MultipleGridBuilder::addGrid), py::arg("grid_shape"))
+ .def("add_grid", py::overload_cast<Object*, uint>(&MultipleGridBuilder::addGrid), py::arg("grid_shape"), py::arg("level_fine"))
+ .def("add_geometry", py::overload_cast<Object*>(&MultipleGridBuilder::addGeometry), py::arg("solid_object"))
+ .def("add_geometry", py::overload_cast<Object*, uint>(&MultipleGridBuilder::addGeometry), py::arg("solid_object"), py::arg("level"))
.def("get_number_of_levels", &MultipleGridBuilder::getNumberOfLevels)
- .def("build_grids", &MultipleGridBuilder::buildGrids);
+ .def("build_grids", &MultipleGridBuilder::buildGrids, py::arg("lbm_or_gks"), py::arg("enable_thin_walls"))
+ .def("set_subdomain_box", &MultipleGridBuilder::setSubDomainBox, py::arg("bounding_box"))
+ .def("find_communication_indices", &MultipleGridBuilder::findCommunicationIndices)
+ .def("set_communication_process", &MultipleGridBuilder::setCommunicationProcess)
+ .def("set_number_of_layers", &MultipleGridBuilder::setNumberOfLayers, py::arg("number_of_layers_fine"), py::arg("number_of_layers_between_levels"));
return gridGeneratorModule;
}
diff --git a/pythonbindings/src/gpu/submodules/grid_provider.cpp b/pythonbindings/src/gpu/submodules/grid_provider.cpp
index 02ff273e2cd1a2022943e19c9a48a447d9dfe54b..717e9d5cd82100636a5398c09662a0895ce8fb56 100644
--- a/pythonbindings/src/gpu/submodules/grid_provider.cpp
+++ b/pythonbindings/src/gpu/submodules/grid_provider.cpp
@@ -1,8 +1,36 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file grid_provider
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include "gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.h"
-// #include <gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.h>
-// #include <gpu/VirtualFluids_GPU/Parameter/Parameter.h>
-// #include "gpu/GridGenerator/grid/GridBuilder/GridBuilder.h"
namespace grid_provider
{
@@ -11,6 +39,6 @@ namespace grid_provider
void makeModule(py::module_ &parentModule)
{
py::class_<GridProvider, std::shared_ptr<GridProvider>>(parentModule, "GridProvider")
- .def("make_grid_generator", &GridProvider::makeGridGenerator, py::return_value_policy::reference);
+ .def_static("make_grid_generator", &GridProvider::makeGridGenerator, py::return_value_policy::reference, py::arg("builder"), py::arg("para"), py::arg("cuda_memory_manager"), py::arg("communicator"));
}
}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/grid_scaling_factory.cpp b/pythonbindings/src/gpu/submodules/grid_scaling_factory.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..a3a572875a4695871c482a4308acab4214dbb481
--- /dev/null
+++ b/pythonbindings/src/gpu/submodules/grid_scaling_factory.cpp
@@ -0,0 +1,52 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file grid_scaling_factory.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
+#include <pybind11/pybind11.h>
+#include <gpu/VirtualFluids_GPU/Factories/GridScalingFactory.h>
+
+namespace grid_scaling_factory
+{
+ namespace py = pybind11;
+
+ void makeModule(py::module_ &parentModule)
+ {
+
+ py::class_<GridScalingFactory, std::shared_ptr<GridScalingFactory>>(parentModule, "GridScalingFactory")
+ .def(py::init<>())
+ .def("set_scaling_factory", &GridScalingFactory::setScalingFactory, py::arg("scaling_type"));
+
+ py::enum_<GridScalingFactory::GridScaling>(parentModule, "GridScaling")
+ .value("ScaleCompressible", GridScalingFactory::GridScaling::ScaleCompressible)
+ .value("ScaleRhoSq", GridScalingFactory::GridScaling::ScaleRhoSq)
+ .value("NotSpecified", GridScalingFactory::GridScaling::NotSpecified);
+ }
+}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/parameter.cpp b/pythonbindings/src/gpu/submodules/parameter.cpp
index 7b4e67f101e3928abbd4262557864ea1d0f45b02..a7c42223e6a5bfa3caa89c0879e4133fc4123ad0 100644
--- a/pythonbindings/src/gpu/submodules/parameter.cpp
+++ b/pythonbindings/src/gpu/submodules/parameter.cpp
@@ -1,10 +1,46 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file parameter.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <pybind11/functional.h>
#include <pybind11/stl.h>
#include <gpu/VirtualFluids_GPU/Parameter/Parameter.h>
+#include "lbm/constants/NumericConstants.h"
#include <basics/config/ConfigurationFile.h>
#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h>
+
+using namespace vf::lbm::constant;
+
namespace parameter
{
namespace py = pybind11;
@@ -13,42 +49,44 @@ namespace parameter
{
py::class_<Parameter, std::shared_ptr<Parameter>>(parentModule, "Parameter")
.def(py::init<
- const vf::basics::ConfigurationFile&,
int,
- int
- >(),
- "config_data",
- "number_of_processes",
- "my_ID")
- .def("set_forcing", &Parameter::setForcing)
- .def("set_diff_on", &Parameter::setDiffOn)
- .def("set_comp_on", &Parameter::setCompOn)
- .def("set_max_level", &Parameter::setMaxLevel)
- .def("set_t_end", &Parameter::setTEnd)
- .def("set_t_out", &Parameter::setTOut)
- .def("set_t_start_out", &Parameter::setTStartOut)
- .def("set_timestep_of_coarse_level", &Parameter::setTimestepOfCoarseLevel)
- .def("set_output_path", &Parameter::setOutputPath)
- .def("set_output_prefix", &Parameter::setOutputPrefix)
- .def("set_f_name", &Parameter::setFName)
- .def("set_print_files", &Parameter::setPrintFiles)
- .def("set_temperature_init", &Parameter::setTemperatureInit)
- .def("set_temperature_BC", &Parameter::setTemperatureBC)
- .def("set_viscosity", &Parameter::setViscosity)
- .def("set_velocity", &Parameter::setVelocity)
- .def("set_viscosity_ratio", &Parameter::setViscosityRatio)
- .def("set_velocity_ratio", &Parameter::setVelocityRatio)
- .def("set_density_ratio", &Parameter::setDensityRatio)
- .def("set_devices", &Parameter::setDevices)
- .def("set_is_body_force", &Parameter::setIsBodyForce)
- .def("set_use_AMD", &Parameter::setUseAMD)
- .def("set_use_Wale", &Parameter::setUseWale)
- .def("set_SGS_constant", &Parameter::setSGSConstant)
- .def("set_main_kernel", &Parameter::setMainKernel)
- .def("set_AD_kernel", &Parameter::setADKernel)
- .def("set_use_AMD", &Parameter::setUseAMD)
- .def("set_use_Wale", &Parameter::setUseWale)
- .def("set_SGS_constant", &Parameter::setSGSConstant)
+ int,
+ std::optional<const vf::basics::ConfigurationFile*>>(),
+ py::arg("number_of_processes"),
+ py::arg("my_ID"),
+ py::arg("config_data"))
+ .def(py::init<int, int>(),
+ py::arg("number_of_processes"),
+ py::arg("my_ID"))
+ .def(py::init<const vf::basics::ConfigurationFile*>(), py::arg("config_data"))
+ .def("set_forcing", &Parameter::setForcing, py::arg("forcing_x"), py::arg("forcing_y"), py::arg("forcing_z"))
+ .def("set_quadric_limiters", &Parameter::setQuadricLimiters, py::arg("quadric_limiter_p"), py::arg("quadric_limiter_m"), py::arg("quadric_limiter_d"))
+ .def("set_diff_on", &Parameter::setDiffOn, py::arg("is_diff"))
+ .def("set_comp_on", &Parameter::setCompOn, py::arg("is_comp"))
+ .def("set_max_level", &Parameter::setMaxLevel, py::arg("number_of_levels"))
+ .def("set_timestep_end", &Parameter::setTimestepEnd, py::arg("tend"))
+ .def("set_timestep_out", &Parameter::setTimestepOut, py::arg("tout"))
+ .def("set_timestep_start_out", &Parameter::setTimestepStartOut, py::arg("t_start_out"))
+ .def("set_timestep_of_coarse_level", &Parameter::setTimestepOfCoarseLevel, py::arg("timestep"))
+ .def("set_calc_turbulence_intensity", &Parameter::setCalcTurbulenceIntensity, py::arg("calc_velocity_and_fluctuations"))
+ .def("set_output_path", &Parameter::setOutputPath, py::arg("o_path"))
+ .def("set_output_prefix", &Parameter::setOutputPrefix, py::arg("o_prefix"))
+ .def("set_print_files", &Parameter::setPrintFiles, py::arg("print_files"))
+ .def("set_temperature_init", &Parameter::setTemperatureInit, py::arg("temp"))
+ .def("set_temperature_BC", &Parameter::setTemperatureBC, py::arg("temp_bc"))
+ .def("set_viscosity_LB", &Parameter::setViscosityLB, py::arg("viscosity"))
+ .def("set_velocity_LB", &Parameter::setVelocityLB, py::arg("velocity"))
+ .def("set_viscosity_ratio", &Parameter::setViscosityRatio, py::arg("viscosity_ratio"))
+ .def("set_velocity_ratio", &Parameter::setVelocityRatio, py::arg("velocity_ratio"))
+ .def("set_density_ratio", &Parameter::setDensityRatio, py::arg("density_ratio"))
+ .def("set_devices", &Parameter::setDevices, py::arg("devices"))
+ .def("set_max_dev", &Parameter::setMaxDev, py::arg("max_dev"))
+ .def("set_is_body_force", &Parameter::setIsBodyForce, py::arg("is_body_force"))
+ .def("set_use_streams", &Parameter::setUseStreams, py::arg("use_streams"))
+ .def("set_main_kernel", &Parameter::setMainKernel, py::arg("kernel"))
+ .def("set_AD_kernel", &Parameter::setADKernel, py::arg("ad_kernel"))
+ .def("set_has_wall_model_monitor", &Parameter::setHasWallModelMonitor, py::arg("has_wall_monitor"))
+ .def("set_outflow_pressure_correction_factor", &Parameter::setOutflowPressureCorrectionFactor, py::arg("correction_factor"))
.def("set_initial_condition", [](Parameter ¶, std::function<std::vector<float>(real, real, real)> &init_func)
{
para.setInitialCondition([init_func](real coordX, real coordY, real coordZ, real& rho, real& vx, real& vy, real& vz)
@@ -59,9 +97,46 @@ namespace parameter
vy = values[2];
vz = values[3];
});
- })
- .def("add_actuator", &Parameter::addActuator)
- .def("add_probe", &Parameter::addProbe)
+ }, py::arg("init_func"))
+ .def("set_initial_condition_uniform", [](Parameter ¶, real velocity_x, real velocity_y, real velocity_z)
+ {
+ para.setInitialCondition([velocity_x, velocity_y, velocity_z](real coordX, real coordY, real coordZ, real& rho, real& vx, real& vy, real& vz) // must capture values explicitly!
+ {
+ rho = c0o1;
+ vx = velocity_x;
+ vy = velocity_y;
+ vz = velocity_z;
+ });
+ }, py::arg("velocity_x"), py::arg("velocity_y"), py::arg("velocity_z"))
+ .def("set_initial_condition_log_law", [](Parameter ¶, real u_star, real z0, real velocityRatio)
+ {
+ para.setInitialCondition(
+ [u_star, z0, velocityRatio](real coordX, real coordY, real coordZ, real& rho, real& vx, real& vy, real& vz)
+ {
+ coordZ = coordZ > c0o1 ? coordZ : c0o1;
+
+ rho = c0o1;
+ vx = u_star/c4o10 * log(coordZ/z0+c1o1) / velocityRatio;
+ vy = c0o1;
+ vz = c0o1;
+ }
+ );
+ }, py::arg("u_star"), py::arg("z0"), py::arg("velocity_ratio"))
+ .def("set_initial_condition_perturbed_log_law", [](Parameter ¶, real u_star, real z0, real L_x, real L_z, real H, real velocityRatio)
+ {
+ para.setInitialCondition(
+ [u_star, z0, L_x, L_z, H, velocityRatio](real coordX, real coordY, real coordZ, real& rho, real& vx, real& vy, real& vz)
+ {
+ coordZ = coordZ > c0o1 ? coordZ : c0o1;
+ rho = c0o1;
+ vx = (u_star/c4o10 * log(coordZ/z0+c1o1) + c2o1*sin(cPi*c16o1*coordX/L_x)*sin(cPi*c8o1*coordZ/H)/(pow(coordZ/H,c2o1)+c1o1)) / velocityRatio;
+ vy = c2o1*sin(cPi*c16o1*coordX/L_x)*sin(cPi*c8o1*coordZ/H)/(pow(coordZ/H,c2o1)+c1o1) / velocityRatio;
+ vz = c8o1*u_star/c4o10*(sin(cPi*c8o1*coordY/H)*sin(cPi*c8o1*coordZ/H)+sin(cPi*c8o1*coordX/L_x))/(pow(c1o2*L_z-coordZ, c2o1)+c1o1) / velocityRatio;
+ }
+ );
+ }, py::arg("u_star"), py::arg("z0"), py::arg("length_x"), py::arg("length_z"), py::arg("height"), py::arg("velocity_ratio"))
+ .def("add_actuator", &Parameter::addActuator, py::arg("actuator"))
+ .def("add_probe", &Parameter::addProbe, py::arg("probe"))
.def("get_output_path", &Parameter::getOutputPath)
.def("get_output_prefix", &Parameter::getOutputPrefix)
.def("get_velocity", &Parameter::getVelocity)
@@ -70,11 +145,9 @@ namespace parameter
.def("get_viscosity_ratio", &Parameter::getViscosityRatio)
.def("get_density_ratio", &Parameter::getDensityRatio)
.def("get_force_ratio", &Parameter::getForceRatio)
- .def("get_use_AMD", &Parameter::getUseAMD)
- .def("get_use_Wale", &Parameter::getUseWale)
.def("get_SGS_constant", &Parameter::getSGSConstant)
.def("get_is_body_force", &Parameter::getIsBodyForce)
- .def("set_has_wall_model_monitor", &Parameter::setHasWallModelMonitor)
;
+
}
}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/pre_collision_interactor.cpp b/pythonbindings/src/gpu/submodules/pre_collision_interactor.cpp
index 362ee1a8ce6112cfa9543f1b254e10f3e35822a1..308f6c37aada14c8c25c69245f603274ae2f18d8 100644
--- a/pythonbindings/src/gpu/submodules/pre_collision_interactor.cpp
+++ b/pythonbindings/src/gpu/submodules/pre_collision_interactor.cpp
@@ -1,3 +1,36 @@
+
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file pre_collision_interactor.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h>
diff --git a/pythonbindings/src/gpu/submodules/precursor_writer.cpp b/pythonbindings/src/gpu/submodules/precursor_writer.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..64164ef9993d7b4f22bff2390b418718f7c3208f
--- /dev/null
+++ b/pythonbindings/src/gpu/submodules/precursor_writer.cpp
@@ -0,0 +1,67 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file precursor_writer.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
+#include <pybind11/pybind11.h>
+#include <pybind11/stl.h>
+#include <pybind11/numpy.h>
+#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h>
+#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/PrecursorWriter.h>
+
+namespace precursor_writer
+{
+ namespace py = pybind11;
+
+ void makeModule(py::module_ &parentModule)
+ {
+ py::enum_<OutputVariable>(parentModule, "OutputVariable")
+ .value("Velocities", OutputVariable::Velocities)
+ .value("Distributions", OutputVariable::Distributions);
+
+ py::class_<PrecursorWriter, PreCollisionInteractor, std::shared_ptr<PrecursorWriter>>(parentModule, "PrecursorWriter")
+ .def(py::init < std::string,
+ std::string,
+ real,
+ real, real,
+ real, real,
+ uint, uint,
+ OutputVariable,
+ uint>(),
+ py::arg("filename"),
+ py::arg("output_path"),
+ py::arg("x_pos"),
+ py::arg("y_min"), py::arg("y_max"),
+ py::arg("z_min"), py::arg("z_max"),
+ py::arg("t_start_out"), py::arg("t_save"),
+ py::arg("output_variable"),
+ py::arg("max_timesteps_per_file"));
+ }
+}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/probes.cpp b/pythonbindings/src/gpu/submodules/probes.cpp
index 6993d9617d870922d7ed90ed9ecbebb8a797be25..7c26958df81a60f00c9909a91f5576a5931652d4 100644
--- a/pythonbindings/src/gpu/submodules/probes.cpp
+++ b/pythonbindings/src/gpu/submodules/probes.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file probes.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <pybind11/stl.h>
#include <gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.h>
@@ -29,7 +61,7 @@ namespace probes
.value("SpatioTemporalFlatness", Statistic::SpatioTemporalFlatness);
py::class_<Probe, PreCollisionInteractor, std::shared_ptr<Probe>>(probeModule, "Probe")
- .def("add_statistic", &Probe::addStatistic)
+ .def("add_statistic", &Probe::addStatistic, py::arg("variable"))
.def("set_file_name_to_n_out", &Probe::setFileNameToNOut)
.def("add_all_available_statistics", &Probe::addAllAvailableStatistics);
@@ -41,14 +73,14 @@ namespace probes
uint,
uint,
uint>(),
- "probe_name",
- "output_path"
- "t_start_avg",
- "t_avg",
- "t_start_out",
- "t_out")
- .def("add_probe_points_from_list", &PointProbe::addProbePointsFromList)
- .def("add_probe_points_from_x_normal_plane", &PointProbe::addProbePointsFromXNormalPlane);
+ py::arg("probe_name"),
+ py::arg("output_path"),
+ py::arg("t_start_avg"),
+ py::arg("t_avg"),
+ py::arg("t_start_out"),
+ py::arg("t_out"))
+ .def("add_probe_points_from_list", &PointProbe::addProbePointsFromList, py::arg("point_coords_x"), py::arg("point_coords_y"), py::arg("point_coords_z"))
+ .def("add_probe_points_from_x_normal_plane", &PointProbe::addProbePointsFromXNormalPlane, py::arg("pos_x"), py::arg("pos0_y"), py::arg("pos0_z"), py::arg("pos1_y"), py::arg("pos1_z"), py::arg("n_y"), py::arg("n_z"));
py::class_<PlaneProbe, Probe, std::shared_ptr<PlaneProbe>>(probeModule, "PlaneProbe")
.def(py::init<
@@ -58,13 +90,13 @@ namespace probes
uint,
uint,
uint>(),
- "probe_name",
- "output_path"
- "t_start_avg",
- "t_avg",
- "t_start_out",
- "t_out")
- .def("set_probe_plane", &PlaneProbe::setProbePlane);
+ py::arg("probe_name"),
+ py::arg("output_path"),
+ py::arg("t_start_avg"),
+ py::arg("t_avg"),
+ py::arg("t_start_out"),
+ py::arg("t_out"))
+ .def("set_probe_plane", &PlaneProbe::setProbePlane, py::arg("pos_x"), py::arg("pos_y"), py::arg("pos_z"), py::arg("delta_x"), py::arg("delta_y"), py::arg("delta_z"));
py::class_<PlanarAverageProbe, Probe, std::shared_ptr<PlanarAverageProbe>>(probeModule, "PlanarAverageProbe")
.def(py::init<
@@ -76,14 +108,14 @@ namespace probes
uint,
uint,
char>(),
- "probe_name",
- "output_path",
- "t_start_avg",
- "t_start_tmp_avg",
- "t_avg",
- "t_start_out",
- "t_out",
- "plane_normal");
+ py::arg("probe_name"),
+ py::arg("output_path"),
+ py::arg("t_start_avg"),
+ py::arg("t_start_tmp_avg"),
+ py::arg("t_avg"),
+ py::arg("t_start_out"),
+ py::arg("t_out"),
+ py::arg("plane_normal"));
py::class_<WallModelProbe, Probe, std::shared_ptr<WallModelProbe>>(probeModule, "WallModelProbe")
@@ -95,15 +127,15 @@ namespace probes
uint,
uint,
uint>(),
- "probe_name",
- "output_path"
- "t_start_avg",
- "t_start_tmp_avg",
- "t_avg",
- "t_start_out",
- "t_out")
- .def("set_force_output_to_stress", &WallModelProbe::setForceOutputToStress)
- .def("set_evaluate_pressure_gradient", &WallModelProbe::setEvaluatePressureGradient);
+ py::arg("probe_name"),
+ py::arg("output_path"),
+ py::arg("t_start_avg"),
+ py::arg("t_start_tmp_avg"),
+ py::arg("t_avg"),
+ py::arg("t_start_out"),
+ py::arg("t_out"))
+ .def("set_force_output_to_stress", &WallModelProbe::setForceOutputToStress, py::arg("output_stress"))
+ .def("set_evaluate_pressure_gradient", &WallModelProbe::setEvaluatePressureGradient, py::arg("eval_press_grad"));
return probeModule;
}
diff --git a/pythonbindings/src/gpu/submodules/simulation.cpp b/pythonbindings/src/gpu/submodules/simulation.cpp
index b775d604ba41530223f22738c72785b2c15348b3..d32ef272a1fd26510439dde6ab3a9438d68009a7 100644
--- a/pythonbindings/src/gpu/submodules/simulation.cpp
+++ b/pythonbindings/src/gpu/submodules/simulation.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file simulation.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <gpu/VirtualFluids_GPU/LBM/Simulation.h>
#include <gpu/VirtualFluids_GPU/Communication/Communicator.h>
@@ -8,6 +40,9 @@
#include <gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.h>
#include <gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.h>
#include <gpu/VirtualFluids_GPU/Output/DataWriter.h>
+#include "gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.h"
+#include "gpu/VirtualFluids_GPU/TurbulenceModels/TurbulenceModelFactory.h"
+#include "gpu/VirtualFluids_GPU/Factories/GridScalingFactory.h"
namespace simulation
{
@@ -20,13 +55,41 @@ namespace simulation
.def(py::init< std::shared_ptr<Parameter>,
std::shared_ptr<CudaMemoryManager>,
vf::gpu::Communicator &,
- GridProvider &>(),
- "parameter",
- "memoryManager",
- "communicator",
- "gridProvider")
+ GridProvider &,
+ BoundaryConditionFactory*,
+ GridScalingFactory*>(),
+ py::arg("parameter"),
+ py::arg("memoryManager"),
+ py::arg("communicator"),
+ py::arg("gridProvider"),
+ py::arg("bcFactory"),
+ py::arg("gridScalingFactory"))
+ .def(py::init< std::shared_ptr<Parameter>,
+ std::shared_ptr<CudaMemoryManager>,
+ vf::gpu::Communicator &,
+ GridProvider &,
+ BoundaryConditionFactory*>(),
+ py::arg("parameter"),
+ py::arg("memoryManager"),
+ py::arg("communicator"),
+ py::arg("gridProvider"),
+ py::arg("bcFactory"))
+ .def(py::init< std::shared_ptr<Parameter>,
+ std::shared_ptr<CudaMemoryManager>,
+ vf::gpu::Communicator &,
+ GridProvider &,
+ BoundaryConditionFactory*,
+ std::shared_ptr<TurbulenceModelFactory>,
+ GridScalingFactory*>(),
+ py::arg("parameter"),
+ py::arg("memoryManager"),
+ py::arg("communicator"),
+ py::arg("gridProvider"),
+ py::arg("bcFactory"),
+ py::arg("tmFactory"),
+ py::arg("gridScalingFactory"))
.def("run", &Simulation::run)
- .def("addKineticEnergyAnalyzer", &Simulation::addKineticEnergyAnalyzer)
- .def("addEnstrophyAnalyzer", &Simulation::addEnstrophyAnalyzer);
+ .def("addKineticEnergyAnalyzer", &Simulation::addKineticEnergyAnalyzer, py::arg("t_analyse"))
+ .def("addEnstrophyAnalyzer", &Simulation::addEnstrophyAnalyzer, py::arg("t_analyse"));
}
}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/transient_bc_setter.cpp b/pythonbindings/src/gpu/submodules/transient_bc_setter.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..89370ef4c1b91a0c8e480e968a1df3bd4fe540ca
--- /dev/null
+++ b/pythonbindings/src/gpu/submodules/transient_bc_setter.cpp
@@ -0,0 +1,52 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file transient_bc_setter.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
+#include <pybind11/pybind11.h>
+#include <gpu/GridGenerator/TransientBCSetter/TransientBCSetter.h>
+
+namespace transient_bc_setter
+{
+ namespace py = pybind11;
+
+ void makeModule(py::module_ &parentModule)
+ {
+ py::enum_<FileType>(parentModule, "FileType")
+ .value("VTK", FileType::VTK);
+
+ parentModule.def("create_file_collection", &createFileCollection, py::arg("prefix"), py::arg("type"));
+
+ py::class_<FileCollection, std::shared_ptr<FileCollection>>(parentModule, "FileCollection");
+
+ py::class_<VTKFileCollection, FileCollection, std::shared_ptr<VTKFileCollection>>(parentModule, "VTKFileCollection")
+ .def(py::init <std::string>(), py::arg("prefix"));
+ }
+}
\ No newline at end of file
diff --git a/pythonbindings/src/gpu/submodules/turbulence_models.cpp b/pythonbindings/src/gpu/submodules/turbulence_models.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..cfbb9e56127fee0cd90a482dde258d8b96389989
--- /dev/null
+++ b/pythonbindings/src/gpu/submodules/turbulence_models.cpp
@@ -0,0 +1,56 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file turbulence_models.cpp
+//! \ingroup submodules
+//! \author Henry Korb
+//=======================================================================================
+#include "pybind11/pybind11.h"
+#include "gpu/VirtualFluids_GPU/TurbulenceModels/TurbulenceModelFactory.h"
+#include "gpu/VirtualFluids_GPU/LBM/LB.h"
+
+namespace turbulence_model
+{
+ namespace py = pybind11;
+
+ void makeModule(py::module_ &parentModule)
+ {
+ py::enum_<TurbulenceModel>(parentModule, "TurbulenceModel")
+ .value("Smagorinsky", TurbulenceModel::Smagorinsky)
+ .value("AMD", TurbulenceModel::AMD)
+ .value("QR", TurbulenceModel::QR)
+ .value("None", TurbulenceModel::None);
+
+ py::class_<TurbulenceModelFactory, std::shared_ptr<TurbulenceModelFactory>>(parentModule, "TurbulenceModelFactory")
+ .def(py::init< std::shared_ptr<Parameter>>(), py::arg("para"))
+ .def("set_turbulence_model", &TurbulenceModelFactory::setTurbulenceModel, py::arg("turbulence_model"))
+ .def("set_model_constant", &TurbulenceModelFactory::setModelConstant, py::arg("model_constant"))
+ .def("read_config_file", &TurbulenceModelFactory::readConfigFile, py::arg("config_data"));
+
+ }
+}
\ No newline at end of file
diff --git a/pythonbindings/src/lbm/lbm.cpp b/pythonbindings/src/lbm/lbm.cpp
index 441b9ff372f4e4513fee58c4a8a1cd78d38582dd..90fd4a71b0101469666936c89974de316e0e2b18 100644
--- a/pythonbindings/src/lbm/lbm.cpp
+++ b/pythonbindings/src/lbm/lbm.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file lbm.cpp
+//! \ingroup lbm
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
namespace lbm
diff --git a/pythonbindings/src/logger/logger.cpp b/pythonbindings/src/logger/logger.cpp
index 82ad3d92760ae38c0eb62b16be726e4eeaca08ac..555b502fa9a56299895de0fa6dd6cfeb66c15024 100644
--- a/pythonbindings/src/logger/logger.cpp
+++ b/pythonbindings/src/logger/logger.cpp
@@ -1,3 +1,35 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file logging.cpp
+//! \ingroup logger
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <logger/Logger.h>
@@ -10,15 +42,15 @@ namespace logging
py::module loggerModule = parentModule.def_submodule("logger");
py::class_<vf::logging::Logger>(loggerModule, "Logger")
- .def("initialize_logger", &vf::logging::Logger::initalizeLogger)
- .def("change_log_path", &vf::logging::Logger::changeLogPath);
+ .def_static("initialize_logger", &vf::logging::Logger::initalizeLogger)
+ .def_static("change_log_path", &vf::logging::Logger::changeLogPath, py::arg("path"));
// use f-strings (f"text {float}") in python for compounded messages
- loggerModule.def("vf_log_trace", [](std::string arg){ VF_LOG_TRACE(arg); });
- loggerModule.def("vf_log_debug", [](std::string arg){ VF_LOG_DEBUG(arg); });
- loggerModule.def("vf_log_info", [](std::string arg){ VF_LOG_INFO(arg); });
- loggerModule.def("vf_log_warning", [](std::string arg){ VF_LOG_WARNING(arg); });
- loggerModule.def("vf_log_critical", [](std::string arg){ VF_LOG_CRITICAL(arg); });
+ loggerModule.def("vf_log_trace", [](std::string message){ VF_LOG_TRACE(message); }, py::arg("message"));
+ loggerModule.def("vf_log_debug", [](std::string message){ VF_LOG_DEBUG(message); }, py::arg("message"));
+ loggerModule.def("vf_log_info", [](std::string message){ VF_LOG_INFO(message); }, py::arg("message"));
+ loggerModule.def("vf_log_warning", [](std::string message){ VF_LOG_WARNING(message); }, py::arg("message"));
+ loggerModule.def("vf_log_critical", [](std::string message){ VF_LOG_CRITICAL(message); }, py::arg("message"));
return loggerModule;
}
diff --git a/pythonbindings/src/muParser.cpp b/pythonbindings/src/muParser.cpp
index 47408c2758fc92991f1be3113d78b8741215b152..eec39de0b72c21aaa924ea805414847aa9de4492 100644
--- a/pythonbindings/src/muParser.cpp
+++ b/pythonbindings/src/muParser.cpp
@@ -1,9 +1,41 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file muParser.cpp
+//! \ingroup src
+//! \author Henry Korb
+//=======================================================================================
#include <pybind11/pybind11.h>
#include <muParser.h>
namespace py = pybind11;
-PYBIND11_MODULE(pymuparser, m) {
+PYBIND11_MODULE(bindings, m) {
py::class_<mu::ParserBase>(m, "_ParserBase");
py::class_<mu::Parser, mu::ParserBase>(m, "Parser")
diff --git a/setup.cfg b/setup.cfg
new file mode 100644
index 0000000000000000000000000000000000000000..f3bc36f29efe64b81916efa04b9d7831a2abeb09
--- /dev/null
+++ b/setup.cfg
@@ -0,0 +1,11 @@
+[metadata]
+name = pyfluids
+description = Python binding for VirtualFluids
+long_description = file: README.md
+long_description_content_type = text/markdown
+platforms = any
+url = https://git.rz.tu-bs.de/irmb/virtualfluids
+version = 0.0.1
+
+[options]
+python_requires = >=3.6
diff --git a/setup.py b/setup.py
index b26e1c13d09447d17f8e9fd6e2cd0d0671595bf3..530431b3775970b5222bc87d32bfb407363f95d6 100644
--- a/setup.py
+++ b/setup.py
@@ -1,137 +1,72 @@
-import os
-import re
import sys
-import platform
-import subprocess
+from pathlib import Path
+from typing import List
-from setuptools import setup, Extension
-from setuptools.command.build_ext import build_ext
-from setuptools.command.install import install
-from setuptools.command.develop import develop
-from distutils.version import LooseVersion
+import skbuild
"""
-Install python wrapper of virtual fluids
-Install GPU backend with option --GPU
-(pass to pip via --install-option="--GPU")
+Install python wrapper of Virtual Fluids
+install via python:
+ python setup.py install
+ set CMAKE Flags via -DBUILD_VF_GPU:BOOL=ON
+ CMAKE flags have to be separated by --
+ example: python setup.py install -- -DBUILD_VF_CPU:BOOL=ON
+or install via pip:
+ pip install .
+ for pip>21:
+ set CMAKE Flags via --config-settings "-DBUILD_VF_GPU=ON"
+ example: pip install . --config-settings="-DBUILD_VF_GPU=ON"
+ each option has to be passed in individually i.e --config-settings="-DOPT1=ON" --config-settings="-DOPT2=OFF"
+ for pip <21:
+ set CMAKE Flags via --global-option ="-DBUILD_VF_GPU=ON"
+ example: pip install . --global-option="-DBUILD_VF_GPU=ON"
"""
-vf_cmake_args = [
- "-DBUILD_VF_PYTHON_BINDINGS=ON",
- "-DCMAKE_CXX_COMPILER_LAUNCHER=ccache",
- "-DCMAKE_CUDA_COMPILER_LAUNCHER=ccache",
- "-DCMAKE_C_COMPILER_LAUNCHER=ccache",
- "-DBUILD_SHARED_LIBS=OFF",
- "-DBUILD_WARNINGS_AS_ERRORS=OFF"
-]
-
-vf_cpu_cmake_args = [
- "-DBUILD_VF_DOUBLE_ACCURACY=ON",
- "-DBUILD_VF_CPU:BOOL=ON",
- "-DBUILD_VF_UNIT_TESTS:BOOL=ON",
- "-DUSE_METIS=ON",
- "-DUSE_MPI=ON"
-]
-
-vf_gpu_cmake_args = [
- "-DBUILD_VF_DOUBLE_ACCURACY=OFF",
- "-DBUILD_VF_GPU:BOOL=ON",
- "-DBUILD_VF_UNIT_TESTS:BOOL=OFF",
-]
-
-GPU = False
-
-class CommandMixin:
- user_options = [
- ('GPU', None, 'compile pyfluids with GPU backend'),
+package_name = "pyfluids"
+target = "python_bindings"
+src_dir = "pythonbindings"
+stub_package = package_name+"-stubs"
+
+stub_dir = Path(src_dir)/stub_package
+
+
+def add_subfiles(dir_path: Path, suffix: str, root_dir: Path) -> List[str]:
+ files = []
+ for f in dir_path.iterdir():
+ if f.is_dir():
+ files.extend(add_subfiles(f, suffix, root_dir))
+ if f.is_file():
+ if f.suffix != suffix:
+ continue
+ files.append(str(f.relative_to(root_dir)))
+ return files
+
+def add_directory(dir_path: Path, suffix: str):
+ return add_subfiles(dir_path, suffix, dir_path)
+
+stub_files = add_directory(stub_dir, ".pyi")
+
+# hack to get config-args for installation with pip>21
+cmake_args = []
+if "config_args" in locals():
+ cmake_args.extend([f"{k}={v}" for k, v in locals()["config_args"].items()])
+
+cmake_args += [
+ f"-DPython3_ROOT_DIR={Path(sys.prefix)}",
+ "-DBUILD_VF_PYTHON_BINDINGS=ON",
+ "-DBUILD_SHARED_LIBS=OFF",
+ "-DBUILD_VF_DOUBLE_ACCURACY=OFF",
+ "-DBUILD_VF_UNIT_TESTS:BOOL=OFF",
+ "-DBUILD_WARNINGS_AS_ERRORS=OFF",
]
- def initialize_options(self):
- super().initialize_options()
- self.GPU = False
-
- def finalize_options(self):
- super().finalize_options()
-
- def run(self):
- global GPU
- GPU = GPU or self.GPU
- super().run()
-
-
-class InstallCommand(CommandMixin, install):
- user_options = getattr(install, 'user_options', []) + CommandMixin.user_options
-
-
-class DevelopCommand(CommandMixin, develop):
- user_options = getattr(develop, 'user_options', []) + CommandMixin.user_options
-
-
-class CMakeExtension(Extension):
- def __init__(self, name, sourcedir=''):
- Extension.__init__(self, name, sources=[])
- self.sourcedir = os.path.abspath(sourcedir)
-
-
-class CMakeBuild(CommandMixin, build_ext):
- user_options = getattr(build_ext, 'user_options', []) + CommandMixin.user_options
-
- def run(self):
- super().run()
- try:
- out = subprocess.check_output(['cmake', '--version'])
- except OSError:
- raise RuntimeError("CMake must be installed to build the following extensions: " +
- ", ".join(e.name for e in self.extensions))
-
- if platform.system() == "Windows":
- cmake_version = LooseVersion(re.search(r'version\s*([\d.]+)', out.decode()).group(1))
- if cmake_version < '3.1.0':
- raise RuntimeError("CMake >= 3.1.0 is required on Windows")
-
- for ext in self.extensions:
- self.build_extension(ext)
-
- def build_extension(self, ext):
- extdir = os.path.abspath(os.path.dirname(self.get_ext_fullpath(ext.name)))
- # required for auto-detection of auxiliary "native" libs
- if not extdir.endswith(os.path.sep):
- extdir += os.path.sep
-
- cmake_args = ['-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=' + extdir,
- '-DPYTHON_EXECUTABLE=' + sys.executable]
-
- cfg = 'Debug' if self.debug else 'Release'
- build_args = ['--config', cfg]
-
- if platform.system() == "Windows":
- cmake_args += ['-DCMAKE_LIBRARY_OUTPUT_DIRECTORY_{}={}'.format(cfg.upper(), extdir)]
- if sys.maxsize > 2**32:
- cmake_args += ['-A', 'x64']
- build_args += ['--', '/m']
- else:
- cmake_args += ['-DCMAKE_BUILD_TYPE=' + cfg]
- build_args += ['--', '-j2']
-
- cmake_args.extend(vf_cmake_args)
- cmake_args.extend(vf_gpu_cmake_args if GPU else vf_cpu_cmake_args)
-
- env = os.environ.copy()
- env['CXXFLAGS'] = '{} -DVERSION_INFO=\\"{}\\"'.format(env.get('CXXFLAGS', ''),
- self.distribution.get_version())
- if not os.path.exists(self.build_temp):
- os.makedirs(self.build_temp)
- cmake_cache_file = self.build_temp+"/CMakeCache.txt"
- if os.path.exists(cmake_cache_file):
- os.remove(cmake_cache_file)
- subprocess.check_call(['cmake', ext.sourcedir] + cmake_args, cwd=self.build_temp, env=env)
- subprocess.check_call(['cmake', '--build', '.'] + build_args, cwd=self.build_temp)
-
-
-setup(
- name='pyfluids',
- version='0.0.1',
- ext_modules=[CMakeExtension('pyfluids')],
- cmdclass={"install": InstallCommand, "develop": DevelopCommand, "build_ext": CMakeBuild},
- zip_safe=False,
+skbuild.setup(
+ name=package_name,
+ packages=[package_name, "pymuparser", "pyfluids-stubs"],
+ package_dir={"": src_dir},
+ cmake_args=cmake_args,
+ cmake_install_target=target,
+ package_data={ "pyfluids": ["py.typed"],
+ "pyfluids-stubs": stub_files},
+ include_package_data=True,
)
diff --git a/src/basics/basics/utilities/UbTuple.h b/src/basics/basics/utilities/UbTuple.h
index fe9c787cead38621beafab3d082122277bdcff73..228ab48898e5e61777d2fcc0061eb6f0434d5cad 100644
--- a/src/basics/basics/utilities/UbTuple.h
+++ b/src/basics/basics/utilities/UbTuple.h
@@ -597,6 +597,8 @@ inline UbTuple<T1, T2, T3, T4, T5, T6, T7, T8> makeUbTuple(T1 const &a1, T2 cons
// some typedefs
using UbTupleFloat2 = UbTuple<float, float>;
using UbTupleFloat3 = UbTuple<float, float, float>;
+using UbTupleFloat4 = UbTuple<float, float, float, float>;
+using UbTupleFloat6 = UbTuple<float, float, float,float, float, float>;
using UbTupleInt2 = UbTuple<int, int>;
using UbTupleInt3 = UbTuple<int, int, int>;
using UbTupleInt4 = UbTuple<int, int, int, int>;
diff --git a/src/basics/basics/writer/WbWriterVtkXmlImageBinary.cpp b/src/basics/basics/writer/WbWriterVtkXmlImageBinary.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..92434e18ba1a2d4b2bc2027c6697f2d1f8393cf5
--- /dev/null
+++ b/src/basics/basics/writer/WbWriterVtkXmlImageBinary.cpp
@@ -0,0 +1,362 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file WbWriterVtkXmlImageBinary.cpp
+//! \ingroup writer
+//! \author Soeren Freudiger, Sebastian Geller, Henry Korb, Henrik Asmuth
+//=======================================================================================
+#include <basics/utilities/UbLogger.h>
+#include <basics/writer/WbWriterVtkXmlImageBinary.h>
+#include <cstring>
+
+using namespace std;
+
+/*===============================================================================*/
+const std::string WbWriterVtkXmlImageBinary::pvdEndTag = " </Collection>\n</VTKFile>";
+/*===============================================================================*/
+string WbWriterVtkXmlImageBinary::writeCollection(const string &filename, const vector<string> &filenames,
+ const double &timeStep, const bool &sepGroups)
+{
+ string vtkfilename = filename + ".pvd";
+ ofstream out(vtkfilename.c_str());
+ if (!out) {
+ out.clear(); // flags ruecksetzen (ansonsten liefert utern if(!out) weiterhin true!!!
+ string path = UbSystem::getPathFromString(vtkfilename);
+ if (path.size() > 0) {
+ UbSystem::makeDirectory(path);
+ out.open(vtkfilename.c_str());
+ }
+ if (!out)
+ throw UbException(UB_EXARGS, "couldn't open file " + vtkfilename);
+ }
+
+ string endian;
+ if (UbSystem::isLittleEndian())
+ endian = "LittleEndian";
+ else
+ endian = "BigEndian";
+ out << "<VTKFile type=\"Collection\" version=\"0.1\" byte_order=\"" << endian << "\" >" << endl;
+ out << " <Collection>" << endl;
+
+ int group = 0, part = 0;
+ for (size_t i = 0; i < filenames.size(); i++) {
+ out << " <DataSet timestep=\"" << timeStep << "\" group=\"" << group << "\" part=\"" << part
+ << "\" file=\"" << filenames[i] << "\"/>" << endl;
+ if (sepGroups)
+ group++;
+ else
+ part++;
+ }
+ out << pvdEndTag;
+ out.close();
+
+ return vtkfilename;
+}
+/*===============================================================================*/
+string WbWriterVtkXmlImageBinary::addFilesToCollection(const string &filename, const vector<string> &filenames,
+ const double &timeStep, const bool &sepGroups)
+{
+ string vtkfilename = filename;
+ fstream test(vtkfilename.c_str(), ios::in);
+ if (!test) {
+ test.clear();
+ vtkfilename += ".pvd";
+ test.open(vtkfilename.c_str(), ios::in);
+ if (!test)
+ return this->writeCollection(filename, filenames, timeStep, sepGroups);
+ }
+
+ fstream out(vtkfilename.c_str(), ios::in | ios::out);
+ out.seekp(-(int)pvdEndTag.size() - 1, ios_base::end);
+
+ int group = 0;
+ for (size_t i = 0; i < filenames.size(); i++) {
+ out << " <DataSet timestep=\"" << timeStep << "\" group=\"" << group << "\" part=\"" << i << "\" file=\""
+ << filenames[i] << "\"/>" << endl;
+ if (sepGroups)
+ group++;
+ }
+ out << pvdEndTag;
+
+ return vtkfilename;
+}
+/*===============================================================================*/
+string WbWriterVtkXmlImageBinary::writeParallelFile(const string &filename, const UbTupleInt6 &wholeExtent, const UbTupleFloat3 &origin, const UbTupleFloat3 &spacing,
+ vector<string> &pieceSources, vector<UbTupleInt6> &pieceExtents,
+ vector<string> &pointDataNames, vector<string> &cellDataNames)
+{
+ string vtkfilename = filename + ".pvti";
+ UBLOG(logDEBUG1, "WbWriterVtkXmlImageBinary::writeParallelFile to " << vtkfilename << " - start");
+
+ ofstream out(vtkfilename.c_str());
+ if (!out) {
+ out.clear(); // flags ruecksetzen (ansonsten liefert utern if(!out) weiterhin true!!!
+ string path = UbSystem::getPathFromString(vtkfilename);
+ if (path.size() > 0) {
+ UbSystem::makeDirectory(path);
+ out.open(vtkfilename.c_str());
+ }
+ if (!out)
+ throw UbException(UB_EXARGS, "couldn't open file " + vtkfilename);
+ }
+
+ // VTK FILE
+ out << "<VTKFile type=\"PImageData\" version=\"0.1\" byte_order=\"LittleEndian\">"
+ << "\n";
+ out << " <PImageData "
+ << "WholeExtent=\"" << val<1>(wholeExtent) << " "
+ << val<2>(wholeExtent) << " "
+ << val<3>(wholeExtent) << " "
+ << val<4>(wholeExtent) << " "
+ << val<5>(wholeExtent) << " "
+ << val<6>(wholeExtent) << "\" "
+ << "GhostLevel=\"0\" "
+ << "Origin=\"" << val<1>(origin) << " "
+ << val<2>(origin) << " "
+ << val<3>(origin) << "\" "
+ << "Spacing=\"" << val<1>(spacing) << " "
+ << val<2>(spacing) << " "
+ << val<3>(spacing) << "\" "
+ << "> \n";
+ out << " <PPointData>\n";
+ for (size_t s = 0; s < pointDataNames.size(); s++)
+ out << " <PDataArray type=\"Float32\" Name=\"" << pointDataNames[s] << "\"/>\n";
+ out << " </PPointData>\n";
+ if (cellDataNames.size() > 0) {
+ out << " <PCellData>\n";
+ for (size_t s = 0; s < cellDataNames.size(); s++)
+ out << " <PDataArray type=\"Float32\" Name=\"" << cellDataNames[s] << "\"/>\n";
+ out << " </PCellData>\n";
+ }
+ for (size_t s = 0; s < pieceSources.size(); s++)
+ out << " <Piece Extent=\"" << val<1>(pieceExtents[s]) << " "
+ << val<2>(pieceExtents[s]) << " "
+ << val<3>(pieceExtents[s]) << " "
+ << val<4>(pieceExtents[s]) << " "
+ << val<5>(pieceExtents[s]) << " "
+ << val<6>(pieceExtents[s]) << "\" Source=\"" << pieceSources[s] << "\"/>\n";
+ out << " </PImageData>\n";
+ out << "</VTKFile>";
+ out << endl;
+ out.close();
+ UBLOG(logDEBUG1, "WbWriterVtkXmlImageBinary::writeParallelFile to " << vtkfilename << " - end");
+
+ return vtkfilename;
+}
+/*===============================================================================*/
+string WbWriterVtkXmlImageBinary::writeOctsWithCellData(const string &filename, vector<UbTupleFloat3> &nodes,
+ vector<UbTupleInt8> &cells, vector<string> &datanames,
+ vector<vector<double>> &celldata)
+{
+ string vtkfilename = filename + getFileExtension();
+ UBLOG(logDEBUG1, "WbWriterVtkXmlImageBinary::writeOctsWithCellData to " << vtkfilename << " - start");
+
+ vector<string> nodeDataNames;
+ vector<vector<double>> nodedata;
+
+ UbTupleFloat3 origin, spacing;
+ UbTupleInt6 extent;
+
+ getMetaDataOfImage(nodes, origin, spacing, extent);
+
+ this->writeData(vtkfilename, nodeDataNames, datanames, nodedata, celldata, extent, origin, spacing, extent);
+ UBLOG(logDEBUG1, "WbWriterVtkXmlImageBinary::writeOctsWithCellData to " << vtkfilename << " - end");
+
+ return vtkfilename;
+}
+/*===============================================================================*/
+string WbWriterVtkXmlImageBinary::writeOctsWithNodeData(const string &filename, vector<UbTupleFloat3> &nodes,
+ vector<UbTupleUInt8> &cells, vector<string> &datanames,
+ vector<vector<double>> &nodedata)
+{
+ string vtkfilename = filename + getFileExtension();
+ UBLOG(logDEBUG1, "WbWriterVtkXmlImageBinary::writeOctsWithNodeData to " << vtkfilename << " - start");
+
+ vector<string> cellDataNames;
+ vector<vector<double>> cellData;
+
+ UbTupleFloat3 origin, spacing;
+ UbTupleInt6 extent;
+
+ getMetaDataOfImage(nodes, origin, spacing, extent);
+
+ this->writeData(vtkfilename, datanames, cellDataNames, nodedata, cellData, extent, origin, spacing, extent);
+
+ UBLOG(logDEBUG1, "WbWriterVtkXmlImageBinary::writeOctsWithNodeData to " << vtkfilename << " - end");
+
+ return vtkfilename;
+}
+/*===============================================================================*/
+string WbWriterVtkXmlImageBinary::writeNodesWithNodeData(const string &filename, vector<UbTupleFloat3> &nodes,
+ vector<string> &datanames,
+ vector<vector<double>> &nodedata)
+{
+ string vtkfilename = filename + getFileExtension();
+ UBLOG(logDEBUG1, "WbWriterVtkXmlImageBinary::writeNodesWithNodeData to " << vtkfilename << " - start");
+
+ vector<string> cellDataNames;
+ vector<vector<double>> cellData;
+
+ UbTupleFloat3 origin, spacing;
+ UbTupleInt6 extent;
+
+ getMetaDataOfImage(nodes, origin, spacing, extent);
+ this->writeData(vtkfilename, datanames, cellDataNames, nodedata, cellData, extent, origin, spacing, extent);
+
+ return vtkfilename;
+}
+
+void WbWriterVtkXmlImageBinary::getMetaDataOfImage(vector<UbTupleFloat3> &nodes, UbTupleFloat3& origin, UbTupleFloat3& spacing, UbTupleInt6& extent)
+{
+ int nofNodes = (int)nodes.size();
+ val<1>(origin) = val<1>(nodes[0]);
+ val<2>(origin) = val<2>(nodes[0]);
+ val<3>(origin) = val<3>(nodes[0]);
+
+ float l_x = val<1>(nodes[nofNodes-1])-val<1>(origin);
+ float l_y = val<2>(nodes[nofNodes-1])-val<2>(origin);
+
+ val<1>(spacing) = val<1>(nodes[1])-val<1>(nodes[0]);
+ int nx = (l_x) / val<1>(spacing);
+ val<2>(spacing) = val<2>(nodes[nx])-val<2>(nodes[0]);
+ int ny = (l_y) / val<2>(spacing);
+ val<3>(spacing) = val<3>(nodes[nx*ny])-val<3>(nodes[0]);
+
+ val<1>(extent) = val<1>(origin)/val<1>(spacing); val<2>(extent) = val<1>(nodes[nofNodes-1])/val<1>(spacing);
+ val<3>(extent) = val<2>(origin)/val<2>(spacing); val<4>(extent) = val<2>(nodes[nofNodes-1])/val<2>(spacing);
+ val<5>(extent) = val<3>(origin)/val<3>(spacing); val<6>(extent) = val<3>(nodes[nofNodes-1])/val<3>(spacing);
+
+}
+
+void WbWriterVtkXmlImageBinary::writeData(const string &vtkfilename,
+ vector<string> &pointDataNames, vector<string> &cellDataNames,
+ vector<vector<double>> &nodedata, vector<vector<double>> &celldata,
+ UbTupleInt6& wholeExtent,
+ UbTupleFloat3& origin, UbTupleFloat3& spacing, UbTupleInt6& extent, unsigned int precision)
+{
+ ofstream out(vtkfilename.c_str(), ios::out | ios::binary);
+ out.precision(precision);
+
+ if (!out) {
+ out.clear(); // flags ruecksetzen (ansonsten liefert utern if(!out) weiterhin true!!!
+ string path = UbSystem::getPathFromString(vtkfilename);
+ if (path.size() > 0) {
+ UbSystem::makeDirectory(path);
+ out.open(vtkfilename.c_str(), ios::out | ios::binary);
+ }
+ if (!out)
+ throw UbException(UB_EXARGS, "couldn't open file " + vtkfilename);
+ }
+
+ size_t nPoints = pointDataNames.size()>0 ? nodedata[0].size() : celldata[0].size();
+
+ int bytesPerByteVal = 4; //==sizeof(int)
+
+ int bytesScalarData = 1 /*scalar */ * (int)nPoints * sizeof(double);
+
+ int offset = 0;
+
+ // VTK FILE
+ out << "<?xml version=\"1.0\"?>\n";
+ out << "<VTKFile type=\"ImageData\" version=\"0.1\" byte_order=\"LittleEndian\" >"
+ << "\n";
+ out << " <ImageData "
+ << "WholeExtent=\"" << val<1>(wholeExtent) << " "
+ << val<2>(wholeExtent) << " "
+ << val<3>(wholeExtent) << " "
+ << val<4>(wholeExtent) << " "
+ << val<5>(wholeExtent) << " "
+ << val<6>(wholeExtent) << "\" "
+ << "Origin=\"" << val<1>(origin) << " "
+ << val<2>(origin) << " "
+ << val<3>(origin) << "\" "
+ << "Spacing=\"" << val<1>(spacing) << " "
+ << val<2>(spacing) << " "
+ << val<3>(spacing) << "\""
+ << "> \n";
+ out << " <Piece Extent=\"" << val<1>(extent) << " "
+ << val<2>(extent) << " "
+ << val<3>(extent) << " "
+ << val<4>(extent) << " "
+ << val<5>(extent) << " "
+ << val<6>(extent) << "\">\n";
+
+ // DATA SECTION
+ if (pointDataNames.size()>0)
+ {
+ out << " <PointData>\n";
+ for (size_t s = 0; s < pointDataNames.size(); ++s) {
+ out << " <DataArray type=\"Float64\" Name=\"" << pointDataNames[s] << "\" format=\"appended\" offset=\""
+ << offset << "\" /> \n";
+ offset += (bytesPerByteVal + bytesScalarData);
+ }
+ out << " </PointData>\n";
+ }
+
+ if (cellDataNames.size()>0)
+ {
+ out << " <CellData>\n";
+ for (size_t s = 0; s < cellDataNames.size(); ++s) {
+ out << " <DataArray type=\"Float64\" Name=\"" << cellDataNames[s] << "\" format=\"appended\" offset=\""
+ << offset << "\" /> \n";
+ offset += (bytesPerByteVal + bytesScalarData);
+ }
+ out << " </CellData>\n";
+ }
+
+ out << " </Piece>\n";
+ out << " </ImageData>\n";
+
+ // AppendedData SECTION
+ out << " <AppendedData encoding=\"raw\">\n";
+ out << "_";
+
+
+ // DATA SECTION
+ // pointData
+ for (size_t s = 0; s < pointDataNames.size(); ++s) {
+ out.write((char *)&bytesScalarData, bytesPerByteVal);
+ for (size_t d = 0; d < nodedata[s].size(); ++d) {
+ double tmp = nodedata[s][d];
+ out.write((char *)&tmp, sizeof(double));
+ }
+ }
+
+ // cellData
+ for (size_t s = 0; s < cellDataNames.size(); ++s) {
+ out.write((char *)&bytesScalarData, bytesPerByteVal);
+ for (size_t d = 0; d < celldata[s].size(); ++d) {
+ double tmp = celldata[s][d];
+ out.write((char *)&tmp, sizeof(double));
+ }
+ }
+ out << "\n </AppendedData>\n";
+ out << "</VTKFile>";
+ out << endl;
+ out.close();
+}
\ No newline at end of file
diff --git a/src/basics/basics/writer/WbWriterVtkXmlImageBinary.h b/src/basics/basics/writer/WbWriterVtkXmlImageBinary.h
new file mode 100644
index 0000000000000000000000000000000000000000..c41ff442732e5f65db0f1dd1ec63e5c3ffca1486
--- /dev/null
+++ b/src/basics/basics/writer/WbWriterVtkXmlImageBinary.h
@@ -0,0 +1,110 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file WbWriterVtkXmlBinary.h
+//! \ingroup writer
+//! \author Soeren Freudiger, Sebastian Geller
+//=======================================================================================
+#ifndef WBWRITERVTKXMLIMAGEBINARY_H
+#define WBWRITERVTKXMLIMAGEBINARY_H
+
+#include <string>
+
+#include <basics/writer/WbWriter.h>
+
+#include "basics_export.h"
+
+class BASICS_EXPORT WbWriterVtkXmlImageBinary : public WbWriter
+{
+public:
+ static WbWriterVtkXmlImageBinary *getInstance()
+ {
+ static WbWriterVtkXmlImageBinary instance;
+ return &instance;
+ }
+
+ WbWriterVtkXmlImageBinary(const WbWriterVtkXmlImageBinary &) = delete;
+ const WbWriterVtkXmlImageBinary &operator=(const WbWriterVtkXmlImageBinary &) = delete;
+
+private:
+ WbWriterVtkXmlImageBinary() : WbWriter()
+ {
+ if (sizeof(unsigned char) != 1)
+ throw UbException(UB_EXARGS, "machine error char type mismatch");
+ if (sizeof(int) != 4)
+ throw UbException(UB_EXARGS, "machine error int type mismatch");
+ if (sizeof(float) != 4)
+ throw UbException(UB_EXARGS, "machine error float type mismatch");
+ }
+
+ static const std::string pvdEndTag;
+
+public:
+ std::string getFileExtension() override { return ".bin.vti"; }
+
+ // write a metafile
+ std::string writeCollection(const std::string &filename, const std::vector<std::string> &filenames,
+ const double ×tep, const bool &sepGroups);
+ std::string addFilesToCollection(const std::string &filename, const std::vector<std::string> &filenames,
+ const double ×tep, const bool &sepGroups);
+ std::string writeParallelFile(const std::string &filename, const UbTupleInt6 &wholeExtent, const UbTupleFloat3 &origin, const UbTupleFloat3 &spacing,
+ std::vector<std::string> &pieceSources, std::vector<UbTupleInt6> &pieceExtents,
+ std::vector<std::string> &pointDataNames, std::vector<std::string> &cellDataNames);
+
+ //////////////////////////////////////////////////////////////////////////
+ // nodes
+ std::string writeNodesWithNodeData(const std::string &filename, std::vector<UbTupleFloat3> &nodes,
+ std::vector<std::string> &datanames,
+ std::vector<std::vector<double>> &nodedata) override;
+
+ //////////////////////////////////////////////////////////////////////////
+ // octs
+ // 7 ---- 6
+ // /| /|
+ // 4 +--- 5 |
+ // | | | |
+ // | 3 ---+ 2
+ // |/ |/
+ // 0 ---- 1
+ std::string writeOctsWithCellData(const std::string &filename, std::vector<UbTupleFloat3> &nodes,
+ std::vector<UbTupleInt8> &cells, std::vector<std::string> &datanames,
+ std::vector<std::vector<double>> &celldata) override;
+ std::string writeOctsWithNodeData(const std::string &filename, std::vector<UbTupleFloat3> &nodes,
+ std::vector<UbTupleUInt8> &cells, std::vector<std::string> &datanames,
+ std::vector<std::vector<double>> &nodedata) override;
+ void writeData(const std::string &vtkfilename,
+ std::vector<std::string> &pointDataNames, std::vector<std::string> &cellDataNames,
+ std::vector<std::vector<double>> &nodedata, std::vector<std::vector<double>> &celldata,
+ UbTupleInt6 &wholeExtent,
+ UbTupleFloat3 &origin, UbTupleFloat3 &spacing, UbTupleInt6 &extent, unsigned int precision=6);
+
+private:
+ void getMetaDataOfImage(std::vector<UbTupleFloat3> &nodes, UbTupleFloat3& origin, UbTupleFloat3& spacing, UbTupleInt6& extent);
+};
+
+#endif // WBWRITERVTKXMLIMAGEBINARY_H
diff --git a/src/basics/config/ConfigurationFile.h b/src/basics/config/ConfigurationFile.h
index ef7e7c9f06f94cabb3ba9cbefe95c8ee75736958..4a53f7add85b9c6461fda0bab20fa6656eebc5d3 100644
--- a/src/basics/config/ConfigurationFile.h
+++ b/src/basics/config/ConfigurationFile.h
@@ -64,6 +64,10 @@ public:
template<class T>
T getValue(const std::string& key) const;
+ //! get value with key and default value
+ template<class T>
+ T getValue(const std::string& key, T defaultValue) const;
+
private:
//! the container
std::map<std::string, std::string> data;
@@ -138,6 +142,19 @@ T ConfigurationFile::getValue(const std::string& key) const
return x;
}
+template<class T>
+T ConfigurationFile::getValue(const std::string& key, T defaultValue) const
+{
+ if (contains(key))
+ {
+ return getValue<T>(key);
+ }
+ else
+ {
+ return defaultValue;
+ }
+}
+
}
#endif
diff --git a/src/gpu/GridGenerator/TransientBCSetter/TransientBCSetter.cpp b/src/gpu/GridGenerator/TransientBCSetter/TransientBCSetter.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..3026458c988c4e7624e15304f144d01f40fb0cdf
--- /dev/null
+++ b/src/gpu/GridGenerator/TransientBCSetter/TransientBCSetter.cpp
@@ -0,0 +1,440 @@
+#include "TransientBCSetter.h"
+#include "GridGenerator/grid/Grid.h"
+#include "GridGenerator/grid/BoundaryConditions/BoundaryCondition.h"
+#include <logger/Logger.h>
+
+
+#include <math.h>
+#include <sstream>
+#include <fstream>
+#include <iostream>
+#include <algorithm>
+
+SPtr<FileCollection> createFileCollection(std::string prefix, FileType type)
+{
+ switch(type)
+ {
+ case FileType::VTK:
+ return std::make_shared<VTKFileCollection>(prefix);
+ break;
+ default:
+ return nullptr;
+ }
+}
+
+SPtr<TransientBCInputFileReader> createReaderForCollection(SPtr<FileCollection> fileCollection, uint readLevel)
+{
+ switch(fileCollection->getFileType())
+ {
+ case FileType::VTK:
+ return std::make_shared<VTKReader>(std::static_pointer_cast<VTKFileCollection>(fileCollection), readLevel);
+ break;
+ default:
+ return nullptr;
+ }
+}
+
+template<typename T>
+std::vector<T> readStringToVector(std::string s)
+{
+ std::vector<T> out;
+ std::stringstream input(s);
+ float num;
+ while(input >> num)
+ {
+ out.push_back(num);
+ }
+ return out;
+}
+
+std::string readElement(std::string line)
+{
+ size_t elemStart = line.find("<")+1;
+ // size_t elemEnd = line.find("/>", elemStart);
+ size_t nameLen = line.find(" ", elemStart)-elemStart;
+ return line.substr(elemStart, nameLen);
+}
+
+std::string readAttribute(std::string line, std::string attributeName)
+{
+ size_t attributeStart = line.find(attributeName)+attributeName.size() + 2; // add 2 for '="'
+ size_t attributeLen = line.find("\"", attributeStart)-attributeStart;
+ return line.substr(attributeStart, attributeLen);
+}
+
+void VTKFile::readHeader()
+{
+ //TODO make this more flexible
+ std::ifstream file(this->fileName);
+
+ std::string line;
+
+ getline(file, line); // VTKFile
+ if(line[1]=='?') getline(file, line); // ignore first line if xml version
+
+ getline(file, line); // ImageData
+ std::vector<int> wholeExtent = readStringToVector<int>(readAttribute(line, "WholeExtent"));
+ std::vector<float> origin = readStringToVector<float>(readAttribute(line, "Origin"));
+ std::vector<float> spacing = readStringToVector<float>(readAttribute(line, "Spacing"));
+
+ getline(file, line); // Piece
+ std::vector<int> pieceExtent = readStringToVector<int>(readAttribute(line, "Extent"));
+ getline(file, line); // PointData
+
+ getline(file, line);
+ while(strcmp(readElement(line).c_str(), "DataArray")==0)
+ {
+ Quantity quant = Quantity();
+ quant.name = readAttribute(line, "Name");
+ quant.offset = std::stoi(readAttribute(line, "offset"));
+ this->quantities.push_back( quant );
+ getline(file, line);
+ }
+ getline(file, line); // </Piece
+ getline(file, line); // </ImageData
+ getline(file, line); // AppendedData
+
+ int offset = int(file.tellg())+sizeof(char)+4; // skip underscore and bytesPerVal
+
+ for(auto& quantity: this->quantities)
+ {
+ quantity.offset += offset;
+ }
+
+ file.close();
+
+ this->deltaX = spacing[0];
+ this->deltaY = spacing[1];
+ this->deltaZ = spacing[2];
+
+ this->nx = pieceExtent[1]-pieceExtent[0]+1;
+ this->ny = pieceExtent[3]-pieceExtent[2]+1;
+ this->nz = pieceExtent[5]-pieceExtent[4]+1;
+
+ this->minX = origin[0]+this->deltaX*pieceExtent[0]; this->maxX = (this->nx-1)*this->deltaX+this->minX;
+ this->minY = origin[1]+this->deltaY*pieceExtent[2]; this->maxY = (this->ny-1)*this->deltaY+this->minY;
+ this->minZ = origin[2]+this->deltaZ*pieceExtent[4]; this->maxZ = (this->nz-1)*this->deltaZ+this->minZ;
+ // printFileInfo();
+
+}
+
+bool VTKFile::markNANs(std::vector<uint> readIndices)
+{
+ std::ifstream buf(fileName.c_str(), std::ios::in | std::ios::binary);
+
+ std::vector<double> tmp;
+ tmp.reserve(readIndices.size());
+ buf.seekg(this->quantities[0].offset);
+ buf.read((char*) tmp.data(), sizeof(double)*readIndices.size());
+ auto firstNAN = std::find_if(tmp.begin(), tmp.end(), [](auto it){ return isnan(it); });
+
+ return firstNAN != tmp.end();
+}
+
+void VTKFile::loadFile()
+{
+ std::ifstream buf(this->fileName.c_str(), std::ios::in | std::ios::binary);
+ for(auto& quantity: this->quantities)
+ {
+ quantity.values.resize(getNumberOfPoints());
+ buf.seekg(quantity.offset);
+ buf.read(reinterpret_cast<char*>(quantity.values.data()), this->getNumberOfPoints()*sizeof(double));
+ }
+
+ buf.close();
+
+ this->loaded = true;
+}
+
+void VTKFile::unloadFile()
+{
+ for(auto& quantity : this->quantities)
+ {
+ std::vector<double> replacement;
+ quantity.values.swap(replacement);
+ }
+ this->loaded = false;
+}
+
+void VTKFile::getData(real* data, uint numberOfNodes, const std::vector<uint>& readIndices, const std::vector<uint>& writeIndices, uint offsetRead, uint offsetWrite)
+{
+ if(!this->loaded) loadFile();
+
+ size_t nPoints = writeIndices.size();
+
+ for(size_t j=0; j<this->quantities.size(); j++)
+ {
+ real* quant = &data[j*numberOfNodes];
+ for(size_t i=0; i<nPoints; i++)
+ {
+ quant[offsetWrite+writeIndices[i]] = this->quantities[j].values[readIndices[i]+offsetRead];
+ }
+ }
+}
+
+void VTKFile::printFileInfo()
+{
+ printf("file %s with \n nx %i ny %i nz %i \n origin %f %f %f \n spacing %f %f %f \n",
+ fileName.c_str(), nx, ny, nz, minX, minY, minZ, deltaX, deltaY, deltaZ);
+ for(auto quantity: this->quantities)
+ {
+ printf("\t quantity %s offset %i \n", quantity.name.c_str(), quantity.offset);
+ }
+
+}
+
+
+void VTKFileCollection::findFiles()
+{
+ bool foundLastLevel = false;
+
+ while(!foundLastLevel)
+ {
+ bool foundLastID = false;
+ std::vector<std::vector<VTKFile>> filesOnThisLevel;
+ while(!foundLastID)
+ {
+ bool foundLastPart = false;
+ std::vector<VTKFile> filesWithThisId;
+ while (!foundLastPart)
+ {
+ std::string fname = makeFileName((int)files.size(), (int)filesOnThisLevel.size(), (int)filesWithThisId.size());
+ std::ifstream f(fname);
+ if(f.good())
+ filesWithThisId.emplace_back(fname);
+ else
+ foundLastPart = true;
+ }
+ if(!filesWithThisId.empty())
+ {
+ VF_LOG_INFO("VTKFileCollection found {} files with ID {} level {}", filesWithThisId.size(), filesOnThisLevel.size(), files.size() );
+ filesOnThisLevel.push_back(filesWithThisId);
+ }
+ else foundLastID = true;
+ }
+
+
+ if(!filesOnThisLevel.empty())
+ files.push_back(filesOnThisLevel);
+ else
+ foundLastLevel = true;
+
+ }
+
+ if(files.empty())
+ VF_LOG_CRITICAL("VTKFileCollection found no files!");
+}
+
+void TransientBCInputFileReader::getNeighbors(uint* neighbor0PP, uint* neighbor0PM, uint* neighbor0MP, uint* neighbor0MM)
+{
+ std::copy(planeNeighbor0PP.begin(), planeNeighbor0PP.end(), &neighbor0PP[writingOffset]);
+ std::copy(planeNeighbor0PM.begin(), planeNeighbor0PM.end(), &neighbor0PM[writingOffset]);
+ std::copy(planeNeighbor0MP.begin(), planeNeighbor0MP.end(), &neighbor0MP[writingOffset]);
+ std::copy(planeNeighbor0MM.begin(), planeNeighbor0MM.end(), &neighbor0MM[writingOffset]);
+}
+
+void TransientBCInputFileReader::getWeights(real* _weights0PP, real* _weights0PM, real* _weights0MP, real* _weights0MM)
+{
+ std::copy(weights0PP.begin(), weights0PP.end(), &_weights0PP[writingOffset]);
+ std::copy(weights0PM.begin(), weights0PM.end(), &_weights0PM[writingOffset]);
+ std::copy(weights0MP.begin(), weights0MP.end(), &_weights0MP[writingOffset]);
+ std::copy(weights0MM.begin(), weights0MM.end(), &_weights0MM[writingOffset]);
+}
+
+
+void VTKReader::initializeIndexVectors()
+{
+ this->readIndices.resize(this->fileCollection->files.size());
+ this->writeIndices.resize(this->fileCollection->files.size());
+ this->nFile.resize(this->fileCollection->files.size());
+ for(size_t lev=0; lev<this->fileCollection->files.size(); lev++)
+ {
+ this->readIndices[lev].resize(this->fileCollection->files[lev].size());
+ this->writeIndices[lev].resize(this->fileCollection->files[lev].size());
+ this->nFile[lev].resize(this->fileCollection->files[lev].size());
+ }
+}
+
+void VTKReader::fillArrays(std::vector<real>& coordsY, std::vector<real>& coordsZ)
+{
+ this->nPoints = (uint)coordsY.size();
+ this->initializeIndexVectors();
+ real max_diff = 1e-4; // maximum distance between point on grid and precursor plane to count as exact match
+ real eps = 1e-7; // small number to avoid division by zero
+ bool perfect_match = true;
+
+ this->weights0PP.reserve(this->nPoints);
+ this->weights0PM.reserve(this->nPoints);
+ this->weights0MP.reserve(this->nPoints);
+ this->weights0MM.reserve(this->nPoints);
+
+ this->planeNeighbor0PP.reserve(this->nPoints);
+ this->planeNeighbor0PM.reserve(this->nPoints);
+ this->planeNeighbor0MP.reserve(this->nPoints);
+ this->planeNeighbor0MM.reserve(this->nPoints);
+
+ for(uint i=0; i<nPoints; i++)
+ {
+
+ real posY = coordsY[i];
+ real posZ = coordsZ[i];
+ bool found0PP = false, found0PM = false, found0MP = false, found0MM = false, foundAll = false;
+
+ uint level = this->readLevel;
+
+ for(int fileId=0; fileId<(int)this->fileCollection->files[level].size(); fileId++)
+ {
+ VTKFile &file = this->fileCollection->files[level][fileId][0];
+ if(!file.inBoundingBox(posY, posZ, 0.0f)) continue;
+
+ // y in simulation is x in precursor/file, z in simulation is y in precursor/file
+ // simulation -> file: N -> E, S -> W, T -> N, B -> S
+ int idx = file.findNeighborMMM(posY, posZ, 0.f); //!> index of nearest WSB neighbor on precursor file
+
+ if(idx!=-1)
+ {
+ // Filter for exact matches
+ if(abs(posY-file.getX(idx)) < max_diff && abs(posZ-file.getY(idx)) < max_diff)
+ {
+ this->weights0PP.emplace_back(1e6f);
+ this->weights0PM.emplace_back(0.f);
+ this->weights0MP.emplace_back(0.f);
+ this->weights0MM.emplace_back(0.f);
+ uint writeIdx = this->getWriteIndex(level, fileId, idx); //!> writeIdx: index on host/device array where precursor value will be written to after loading from file
+ this->planeNeighbor0PP.push_back(writeIdx); //!> neighbor lists mapping where BC kernel should read from on host/device array
+ this->planeNeighbor0PM.push_back(writeIdx);
+ this->planeNeighbor0MP.push_back(writeIdx);
+ this->planeNeighbor0MM.push_back(writeIdx);
+ found0PP = true; found0PM = true; found0MM = true; found0MP = true;
+ }
+ else
+ {
+ perfect_match = false;
+ }
+
+ if(!found0MM)
+ {
+ found0MM = true;
+ real dy = file.getX(idx)-posY;
+ real dz = file.getY(idx)-posZ;
+ this->weights0MM.emplace_back(1.f/(dy*dy+dz*dz+eps));
+ this->planeNeighbor0MM.emplace_back(getWriteIndex(level, fileId, idx));
+ }
+
+ }
+
+ if(!found0PP) //NT in simulation is EN in precursor
+ {
+ int idx = file.findNeighborPPM(posY, posZ, 0.f);
+ if(idx!=-1)
+ {
+ found0PP = true;
+ real dy = file.getX(idx)-posY;
+ real dz = file.getY(idx)-posZ;
+ this->weights0PP.emplace_back(1.f/(dy*dy+dz*dz+eps));
+ this->planeNeighbor0PP.emplace_back(getWriteIndex(level, fileId, idx));
+ }
+ }
+
+ if(!found0PM) //NB in simulation is ES in precursor
+ {
+ int idx = file.findNeighborPMM(posY, posZ, 0.f);
+ if(idx!=-1)
+ {
+ found0PM = true;
+ real dy = file.getX(idx)-posY;
+ real dz = file.getY(idx)-posZ;
+ this->weights0PM.emplace_back(1.f/(dy*dy+dz*dz+eps));
+ this->planeNeighbor0PP.emplace_back(getWriteIndex(level, fileId, idx));
+ }
+ }
+
+ if(!found0MP) //ST in simulation is WN in precursor
+ {
+ int idx = file.findNeighborMPM(posY, posZ, 0.f);
+ if(idx!=-1)
+ {
+ found0MP = true;
+ real dy = file.getX(idx)-posY;
+ real dz = file.getY(idx)-posZ;
+ this->weights0MP.emplace_back(1.f/(dy*dy+dz*dz+eps));
+ this->planeNeighbor0MP.emplace_back(getWriteIndex(level, fileId, idx));
+ }
+ }
+
+ foundAll = found0PP && found0PM && found0MP && found0MM;
+
+ if(foundAll) break;
+ }
+
+ if(!foundAll)
+ {
+ VF_LOG_CRITICAL("Found no matching precursor neighbors for grid point at y={}, z={} \n", posY, posZ);
+ throw std::runtime_error("VTKReader::fillArrays(): Did not find neighbors in the FileCollection for all points");
+ }
+ }
+
+ if(perfect_match)
+ printf("Precursor was a perfect match \n");
+
+
+ for(size_t level=0; level<this->fileCollection->files.size(); level++){
+ for(size_t id=0; id<this->fileCollection->files[level].size(); id++){
+ if(this->fileCollection->files[level][id][0].markNANs(this->readIndices[level][id]))
+ throw std::runtime_error("Found a NAN in the precursor where a velocity is needed");
+ }}
+}
+
+uint VTKReader::getWriteIndex(int level, int id, int linearIndex)
+{
+ auto it = std::find(this->writeIndices[level][id].begin(), this->writeIndices[level][id].end(), linearIndex);
+ uint idx = it-this->writeIndices[level][id].begin();
+ if(it==this->writeIndices[level][id].end())
+ {
+ this->writeIndices[level][id].push_back(this->nPointsRead); //!> index on host/device array where value from file will be written to
+ this->readIndices[level][id].push_back(linearIndex); //!> index in file that will be read from
+ this->nPointsRead++;
+ }
+ return idx;
+}
+
+
+void VTKReader::getNextData(real* data, uint numberOfNodes, real time)
+{
+ // for(size_t level=0; level<this->fileCollection->files.size(); level++)
+ // {
+ uint level = this->readLevel;
+ for(size_t id=0; id<this->fileCollection->files[level].size(); id++)
+ {
+ size_t numberOfFiles = this->nFile[level][id];
+
+
+ if(!this->fileCollection->files[level][id][numberOfFiles].inZBounds(time))
+ {
+ numberOfFiles++;
+
+ printf("switching to precursor file no. %zd\n", numberOfFiles);
+ if(numberOfFiles == this->fileCollection->files[level][id].size())
+ throw std::runtime_error("Not enough Precursor Files to read");
+
+ this->fileCollection->files[level][id][numberOfFiles-1].unloadFile();
+ if(numberOfFiles+1<this->fileCollection->files[level][id].size())
+ {
+ VTKFile* nextFile = &this->fileCollection->files[level][id][numberOfFiles+1];
+ if(! nextFile->isLoaded())
+ {
+ read.wait();
+ read = std::async(std::launch::async, [](VTKFile* file){ file->loadFile(); }, &this->fileCollection->files[level][id][numberOfFiles+1]);
+ }
+ }
+ }
+
+
+ VTKFile* file = &this->fileCollection->files[level][id][numberOfFiles];
+
+ int off = file->getClosestIdxZ(time)*file->getNumberOfPointsInXYPlane();
+ file->getData(data, numberOfNodes, this->readIndices[level][id], this->writeIndices[level][id], off, this->writingOffset);
+ this->nFile[level][id] = numberOfFiles;
+ }
+ // }
+}
\ No newline at end of file
diff --git a/src/gpu/GridGenerator/TransientBCSetter/TransientBCSetter.h b/src/gpu/GridGenerator/TransientBCSetter/TransientBCSetter.h
new file mode 100644
index 0000000000000000000000000000000000000000..5bee61e194670c74ca8bd8da87f3881956fff466
--- /dev/null
+++ b/src/gpu/GridGenerator/TransientBCSetter/TransientBCSetter.h
@@ -0,0 +1,201 @@
+#ifndef TRANSIENTBCSETTER_H_
+#define TRANSIENTBCSETTER_H_
+
+#include "Core/DataTypes.h"
+#include <Core/StringUtilities/StringUtil.h>
+#include "PointerDefinitions.h"
+
+#include <string>
+#include <vector>
+#include <math.h>
+#include <sstream>
+#include <future>
+class Grid;
+namespace gg
+{
+ class BoundaryCondition;
+}
+
+
+enum class FileType
+{
+ VTK
+};
+
+struct Quantity
+{
+ std::string name;
+ int offset;
+ std::vector<double> values;
+};
+
+class VTKFile
+{
+public:
+ VTKFile(std::string _fileName):
+ fileName(_fileName)
+ {
+ readHeader();
+ this->loaded = false;
+ // printFileInfo();
+ };
+
+ void getData(real* data, uint numberOfNodes, const std::vector<uint>& readIndices, const std::vector<uint>& writeIndices, uint offsetRead, uint offsetWrite);
+ bool markNANs(std::vector<uint> readIndices);
+ bool inBoundingBox(real posX, real posY, real posZ){return inXBounds(posX) && inYBounds(posY) && inZBounds(posZ); };
+ bool inXBounds(real posX){ return posX<=maxX && posX>=minX; };
+ bool inYBounds(real posY){ return posY<=maxY && posY>=minY; };
+ bool inZBounds(real posZ){ return posZ<=maxZ && posZ>=minZ; };
+ int findNeighborMMM(real posX, real posY, real posZ){ int idx = getLinearIndex(getIdxM00(posX) , getIdx0M0(posY) , getIdx00M(posZ) ); return (idx>=0) && (idx<nx*ny*nz) ? idx : -1; };
+ int findNeighborMMP(real posX, real posY, real posZ){ int idx = getLinearIndex(getIdxM00(posX) , getIdx0M0(posY) , getIdx00M(posZ)+1); return (idx>=0) && (idx<nx*ny*nz) ? idx : -1; };
+ int findNeighborMPM(real posX, real posY, real posZ){ int idx = getLinearIndex(getIdxM00(posX) , getIdx0M0(posY)+1, getIdx00M(posZ) ); return (idx>=0) && (idx<nx*ny*nz) ? idx : -1; };
+ int findNeighborMPP(real posX, real posY, real posZ){ int idx = getLinearIndex(getIdxM00(posX) , getIdx0M0(posY)+1, getIdx00M(posZ)+1); return (idx>=0) && (idx<nx*ny*nz) ? idx : -1; };
+ int findNeighborPMM(real posX, real posY, real posZ){ int idx = getLinearIndex(getIdxM00(posX)+1, getIdx0M0(posY) , getIdx00M(posZ) ); return (idx>=0) && (idx<nx*ny*nz) ? idx : -1; };
+ int findNeighborPMP(real posX, real posY, real posZ){ int idx = getLinearIndex(getIdxM00(posX)+1, getIdx0M0(posY) , getIdx00M(posZ)+1); return (idx>=0) && (idx<nx*ny*nz) ? idx : -1; };
+ int findNeighborPPM(real posX, real posY, real posZ){ int idx = getLinearIndex(getIdxM00(posX)+1, getIdx0M0(posY)+1, getIdx00M(posZ) ); return (idx>=0) && (idx<nx*ny*nz) ? idx : -1; };
+ int findNeighborPPP(real posX, real posY, real posZ){ int idx = getLinearIndex(getIdxM00(posX)+1, getIdx0M0(posY)+1, getIdx00M(posZ)+1); return (idx>=0) && (idx<nx*ny*nz) ? idx : -1; };
+ int getIdxX(int linearIdx){ return linearIdx%nx;};
+ int getIdxY(int linearIdx){ return (linearIdx/nx)%ny;};
+ int getIdxZ(int linearIdx){ return linearIdx/(nx*ny); };
+ real getX(int linearIdx){ return getIdxX(linearIdx)*deltaX+minX; };
+ real getY(int linearIdx){ return getIdxY(linearIdx)*deltaY+minY; };
+ real getZ(int linearIdx){ return getIdxZ(linearIdx)*deltaZ+minZ; };
+ int getIdxM00(real posX){ return (posX-minX)/deltaX; };
+ int getIdx0M0(real posY){ return (posY-minY)/deltaY; };
+ int getIdx00M(real posZ){ return (posZ-minZ)/deltaZ; };
+ int getClosestIdxX(real posX){ int x = round((posX-minX)/deltaX); return x>nx ? nx : (x<0 ? 0 : x);};
+ int getClosestIdxY(real posY){ int y = round((posY-minY)/deltaY); return y>ny ? ny : (y<0 ? 0 : y);};
+ int getClosestIdxZ(real posZ){ int z = round((posZ-minZ)/deltaZ); return z>nz ? nz : (z<0 ? 0 : z);};
+ int getLinearIndex(int idxX, int idxY, int idxZ){ return idxX + nx*(idxY+ny*idxZ); };
+ int getNumberOfPointsInXYPlane(){ return nx*ny; }
+ int getNumberOfPointsInYZPlane(){ return ny*nz; }
+ int getNumberOfPointsInXZPlane(){ return nx*nz; }
+ int getNumberOfPoints(){ return nx*ny*nz; }
+ size_t getNumberOfQuantities(){ return quantities.size(); }
+ void loadFile();
+ void unloadFile();
+ bool isLoaded(){return loaded;};
+
+
+private:
+ void readHeader();
+ void printFileInfo();
+
+public:
+
+private:
+ std::string fileName;
+ real minX, maxX, minY, maxY, minZ, maxZ;
+ real deltaX, deltaY, deltaZ;
+ int nx, ny, nz;
+ std::vector<Quantity> quantities;
+ bool loaded;
+};
+
+class FileCollection
+{
+public:
+ FileCollection(std::string _prefix):
+ prefix(_prefix){};
+
+ virtual ~FileCollection() = default;
+
+ virtual size_t getNumberOfQuantities()=0;
+
+ virtual FileType getFileType()=0;
+
+protected:
+ std::string prefix;
+};
+
+
+class VTKFileCollection : public FileCollection
+{
+public:
+ VTKFileCollection(std::string _prefix):
+ FileCollection(_prefix)
+ {
+ findFiles();
+ };
+
+ FileType getFileType(){ return FileType::VTK; };
+ size_t getNumberOfQuantities(){ return files[0][0][0].getNumberOfQuantities(); }
+
+
+private:
+ void findFiles();
+ std::string makeFileName(int level, int id, int part)
+ {
+ return prefix + "_lev_" + StringUtil::toString<int>(level)
+ + "_ID_" + StringUtil::toString<int>(id)
+ + "_File_" + StringUtil::toString<int>(part)
+ + ".bin." + suffix;
+ };
+
+
+public:
+ static const inline std::string suffix = "vti";
+ std::vector<std::vector<std::vector<VTKFile>>> files;
+};
+
+
+class TransientBCInputFileReader
+{
+public:
+ TransientBCInputFileReader()
+ {
+ this->nPoints = 0;
+ this->nPointsRead = 0;
+ this->writingOffset = 0;
+ };
+ virtual ~TransientBCInputFileReader() = default;
+
+ virtual void getNextData(real* data, uint numberOfNodes, real time)=0;
+ virtual void fillArrays(std::vector<real>& coordsY, std::vector<real>& coordsZ)=0;
+ uint getNPoints(){return nPoints; };
+ uint getNPointsRead(){return nPointsRead; };
+ size_t getNumberOfQuantities(){ return nQuantities; };
+ void setWritingOffset(uint offset){ this->writingOffset = offset; }
+ void getNeighbors(uint* neighbor0PP, uint* neighbor0PM, uint* neighbor0MP, uint* neighbor0MM);
+ void getWeights(real* _weights0PP, real* _weights0PM, real* _weights0MP, real* _weights0MM);
+
+public:
+ std::vector<uint> planeNeighbor0PP, planeNeighbor0PM, planeNeighbor0MP, planeNeighbor0MM;
+ std::vector<real> weights0PP, weights0PM, weights0MP, weights0MM;
+
+protected:
+ uint nPoints, nPointsRead, writingOffset;
+ uint nReads=0;
+ size_t nQuantities=0;
+};
+
+
+class VTKReader : public TransientBCInputFileReader
+{
+public:
+ VTKReader(SPtr<VTKFileCollection> _fileCollection, uint _readLevel):
+ fileCollection(_fileCollection),
+ readLevel(_readLevel)
+ {
+ this->nQuantities = fileCollection->getNumberOfQuantities();
+ read = std::async([](){});
+ };
+ void getNextData(real* data, uint numberOfNodes, real time) override;
+ void fillArrays(std::vector<real>& coordsY, std::vector<real>& coordsZ) override;
+private:
+ uint getWriteIndex(int level, int id, int linearIdx);
+ void initializeIndexVectors();
+
+private:
+ std::vector<std::vector<std::vector<uint>>> readIndices, writeIndices;
+ std::vector<std::vector<size_t>> nFile;
+ SPtr<VTKFileCollection> fileCollection;
+ uint readLevel;
+ std::future<void> read;
+};
+
+
+SPtr<FileCollection> createFileCollection(std::string prefix, FileType type);
+SPtr<TransientBCInputFileReader> createReaderForCollection(SPtr<FileCollection> fileCollection, uint readLevel);
+
+#endif //TRANSIENTBCSETTER_H_
\ No newline at end of file
diff --git a/src/gpu/GridGenerator/grid/BoundaryConditions/BoundaryCondition.cpp b/src/gpu/GridGenerator/grid/BoundaryConditions/BoundaryCondition.cpp
index 5102f60fc295aadf4323a4b332bf3dd8f7f21dbf..b0fb2604946b83ead45c30adabbcfe8dc26fa656 100644
--- a/src/gpu/GridGenerator/grid/BoundaryConditions/BoundaryCondition.cpp
+++ b/src/gpu/GridGenerator/grid/BoundaryConditions/BoundaryCondition.cpp
@@ -36,12 +36,12 @@
#include "grid/BoundaryConditions/Side.h"
#include "grid/Grid.h"
+#include "GridGenerator/TransientBCSetter/TransientBCSetter.h"
bool gg::BoundaryCondition::isSide( SideType side ) const
{
return this->side->whoAmI() == side;
}
-
//////////////////////////////////////////////////////////////////////////
void VelocityBoundaryCondition::setVelocityProfile(
@@ -124,5 +124,4 @@ void StressBoundaryCondition::fillSamplingIndices(std::vector<SPtr<Grid> > grid,
this->velocitySamplingIndices.push_back( grid[level]->transCoordToIndex(x_sampling, y_sampling, z_sampling) );
}
-}
-
+}
\ No newline at end of file
diff --git a/src/gpu/GridGenerator/grid/BoundaryConditions/BoundaryCondition.h b/src/gpu/GridGenerator/grid/BoundaryConditions/BoundaryCondition.h
index 4a3990d9f815042297be76ae83a61268c8ad6815..22342aec9839afad9bb37b1b11812f6d1750ed7b 100644
--- a/src/gpu/GridGenerator/grid/BoundaryConditions/BoundaryCondition.h
+++ b/src/gpu/GridGenerator/grid/BoundaryConditions/BoundaryCondition.h
@@ -45,6 +45,8 @@ class Grid;
class Side;
enum class SideType;
+class TransientBCInputFileReader;
+
namespace gg
{
class BoundaryCondition
@@ -63,6 +65,8 @@ public:
bool isSide(SideType side) const;
real getQ(uint index, uint dir) { return this->qs[index][dir]; }
+
+ void getCoords( SPtr<Grid> grid, std::vector<real>& x, std::vector<real>& y, std::vector<real>& z);
};
}
@@ -246,6 +250,7 @@ public:
real getVy(uint index) { return this->vyList[index]; }
real getVz(uint index) { return this->vzList[index]; }
+
void setVelocityProfile( SPtr<Grid> grid, std::function<void(real,real,real,real&,real&,real&)> velocityProfile );
};
@@ -329,5 +334,32 @@ public:
real getNormalz(uint index) { return this->normalZList[index]; }
};
+class PrecursorBoundaryCondition : public gg::BoundaryCondition
+{
+public:
+ static SPtr<PrecursorBoundaryCondition> make(SPtr<TransientBCInputFileReader> reader, int timeStepsBetweenReads, real velocityX, real velocityY, real velocityZ)
+ {
+ return SPtr<PrecursorBoundaryCondition>(new PrecursorBoundaryCondition(reader, timeStepsBetweenReads, velocityX, velocityY, velocityZ));
+ }
+ SPtr<TransientBCInputFileReader> getReader(){ return reader; }
+ real getVelocityX() { return velocityX; }
+ real getVelocityY() { return velocityY; }
+ real getVelocityZ() { return velocityZ; }
+
+private:
+ PrecursorBoundaryCondition(SPtr<TransientBCInputFileReader> _reader, uint _timeStepsBetweenReads, real vx, real vy, real vz) : reader(_reader), timeStepsBetweenReads(_timeStepsBetweenReads), velocityX(vx), velocityY(vy), velocityZ(vz) { };
+ virtual char getType() const override
+ {
+ return vf::gpu::BC_VELOCITY;
+ }
+public:
+ uint timeStepsBetweenReads; //!> read data every nth timestep
+
+private:
+ real velocityX = 0.0;
+ real velocityY = 0.0;
+ real velocityZ = 0.0;
+ SPtr<TransientBCInputFileReader> reader;
+};
#endif
\ No newline at end of file
diff --git a/src/gpu/GridGenerator/grid/BoundaryConditions/Side.cpp b/src/gpu/GridGenerator/grid/BoundaryConditions/Side.cpp
index 6c7bf8ca1853826d83fb6a713ffe03716bd2cf9a..0900c2d587ba9811c480427b833e7e083216cf10 100644
--- a/src/gpu/GridGenerator/grid/BoundaryConditions/Side.cpp
+++ b/src/gpu/GridGenerator/grid/BoundaryConditions/Side.cpp
@@ -40,6 +40,18 @@
using namespace gg;
+std::vector<real> Side::getNormal()
+{
+ std::vector<real> normal;
+ if(this->getCoordinate()==X_INDEX)
+ normal = {(real)this->getDirection(), 0.0, 0.0};
+ if(this->getCoordinate()==Y_INDEX)
+ normal = {0.0, (real)this->getDirection(), 0.0};
+ if(this->getCoordinate()==Z_INDEX)
+ normal = {0.0, 0.0, (real)this->getDirection()};
+ return normal;
+}
+
void Side::addIndices(SPtr<Grid> grid, SPtr<BoundaryCondition> boundaryCondition, std::string coord, real constant,
real startInner, real endInner, real startOuter, real endOuter)
{
@@ -49,11 +61,19 @@ void Side::addIndices(SPtr<Grid> grid, SPtr<BoundaryCondition> boundaryCondition
{
const uint index = getIndex(grid, coord, constant, v1, v2);
- if ((index != INVALID_INDEX) && ( grid->getFieldEntry(index) == vf::gpu::FLUID
- || grid->getFieldEntry(index) == vf::gpu::FLUID_CFC
- || grid->getFieldEntry(index) == vf::gpu::FLUID_CFF
- || grid->getFieldEntry(index) == vf::gpu::FLUID_FCC
- || grid->getFieldEntry(index) == vf::gpu::FLUID_FCF ))
+ if ((index != INVALID_INDEX) && ( grid->getFieldEntry(index) == vf::gpu::FLUID
+ || grid->getFieldEntry(index) == vf::gpu::FLUID_CFC
+ || grid->getFieldEntry(index) == vf::gpu::FLUID_CFF
+ || grid->getFieldEntry(index) == vf::gpu::FLUID_FCC
+ || grid->getFieldEntry(index) == vf::gpu::FLUID_FCF
+ || grid->getFieldEntry(index) == vf::gpu::FLUID_FCF
+
+ //! Enforce overlap of BCs on edge nodes
+ || grid->getFieldEntry(index) == vf::gpu::BC_PRESSURE
+ || grid->getFieldEntry(index) == vf::gpu::BC_VELOCITY
+ || grid->getFieldEntry(index) == vf::gpu::BC_NOSLIP
+ || grid->getFieldEntry(index) == vf::gpu::BC_SLIP
+ || grid->getFieldEntry(index) == vf::gpu::BC_STRESS ))
{
grid->setFieldEntry(index, boundaryCondition->getType());
boundaryCondition->indices.push_back(index);
@@ -64,7 +84,6 @@ void Side::addIndices(SPtr<Grid> grid, SPtr<BoundaryCondition> boundaryCondition
boundaryCondition->patches.push_back(0);
}
-
}
}
}
@@ -152,16 +171,21 @@ void Side::setQs(SPtr<Grid> grid, SPtr<BoundaryCondition> boundaryCondition, uin
else neighborZ = grid->getLastFluidNode ( coords, 2, grid->getEndZ() );
}
+ //! Only seting q's that partially point in the Side-normal direction
+ bool alignedWithNormal = (this->getNormal()[0]*grid->getDirection()[dir * DIMENSION + 0]+
+ this->getNormal()[1]*grid->getDirection()[dir * DIMENSION + 1]+
+ this->getNormal()[2]*grid->getDirection()[dir * DIMENSION + 2] ) > 0;
+
uint neighborIndex = grid->transCoordToIndex( neighborX, neighborY, neighborZ );
- if( grid->getFieldEntry(neighborIndex) == vf::gpu::STOPPER_OUT_OF_GRID_BOUNDARY ||
- grid->getFieldEntry(neighborIndex) == vf::gpu::STOPPER_OUT_OF_GRID ||
- grid->getFieldEntry(neighborIndex) == vf::gpu::STOPPER_SOLID )
+ if((grid->getFieldEntry(neighborIndex) == vf::gpu::STOPPER_OUT_OF_GRID_BOUNDARY ||
+ grid->getFieldEntry(neighborIndex) == vf::gpu::STOPPER_OUT_OF_GRID ||
+ grid->getFieldEntry(neighborIndex) == vf::gpu::STOPPER_SOLID) &&
+ alignedWithNormal )
qNode[dir] = 0.5;
else
qNode[dir] = -1.0;
-
}
-
+
boundaryCondition->qs.push_back(qNode);
}
@@ -260,7 +284,7 @@ void MY::addIndices(std::vector<SPtr<Grid> > grid, uint level, SPtr<BoundaryCond
real coordinateNormal = grid[level]->getStartY() + grid[level]->getDelta();
if( coordinateNormal > grid[0]->getStartY() + grid[0]->getDelta() ) return;
-
+
Side::addIndices(grid[level], boundaryCondition, "y", coordinateNormal, startInner, endInner, startOuter, endOuter);
}
diff --git a/src/gpu/GridGenerator/grid/BoundaryConditions/Side.h b/src/gpu/GridGenerator/grid/BoundaryConditions/Side.h
index 6df6bfccc9a39b80de3ac43d057a03945d035b34..53a763bc562ee978042b28d24856fbcca256c5f9 100644
--- a/src/gpu/GridGenerator/grid/BoundaryConditions/Side.h
+++ b/src/gpu/GridGenerator/grid/BoundaryConditions/Side.h
@@ -72,15 +72,17 @@ public:
virtual SideType whoAmI() const = 0;
+ std::vector<real> getNormal();
+
protected:
- static void addIndices(SPtr<Grid> grid, SPtr<gg::BoundaryCondition> boundaryCondition, std::string coord, real constant,
+ void addIndices(SPtr<Grid> grid, SPtr<gg::BoundaryCondition> boundaryCondition, std::string coord, real constant,
real startInner, real endInner, real startOuter, real endOuter);
static void setPressureNeighborIndices(SPtr<gg::BoundaryCondition> boundaryCondition, SPtr<Grid> grid, const uint index);
static void setStressSamplingIndices(SPtr<gg::BoundaryCondition> boundaryCondition, SPtr<Grid> grid, const uint index);
- static void setQs(SPtr<Grid> grid, SPtr<gg::BoundaryCondition> boundaryCondition, uint index);
+ void setQs(SPtr<Grid> grid, SPtr<gg::BoundaryCondition> boundaryCondition, uint index);
private:
static uint getIndex(SPtr<Grid> grid, std::string coord, real constant, real v1, real v2);
diff --git a/src/gpu/GridGenerator/grid/Grid.h b/src/gpu/GridGenerator/grid/Grid.h
index 3f28120a5d969fcc5d7b2a3402a2169ff97c0cc3..85b19bedd470c9856954a1ca20fb446c3d875da2 100644
--- a/src/gpu/GridGenerator/grid/Grid.h
+++ b/src/gpu/GridGenerator/grid/Grid.h
@@ -178,6 +178,21 @@ public:
virtual void findFluidNodeIndicesBorder() = 0;
virtual uint getNumberOfFluidNodesBorder() const = 0;
virtual void getFluidNodeIndicesBorder(uint *fluidNodeIndicesBorder) const = 0;
+
+ virtual void addFluidNodeIndicesMacroVars(std::vector<uint> _fluidNodeIndicesMacroVars) = 0;
+ virtual void addFluidNodeIndicesApplyBodyForce(std::vector<uint> _fluidNodeIndicesApplyBodyForce) = 0;
+ virtual void addFluidNodeIndicesAllFeatures(std::vector<uint> _fluidNodeIndicesAllFeatures) = 0;
+ virtual void sortFluidNodeIndicesMacroVars() = 0;
+ virtual void sortFluidNodeIndicesApplyBodyForce() = 0;
+ virtual void sortFluidNodeIndicesAllFeatures() = 0;
+
+ virtual uint getNumberOfFluidNodeIndicesMacroVars() const = 0;
+ virtual uint getNumberOfFluidNodeIndicesApplyBodyForce() const = 0;
+ virtual uint getNumberOfFluidNodeIndicesAllFeatures() const = 0;
+ virtual void getFluidNodeIndicesMacroVars(uint *fluidNodeIndicesMacroVars) const = 0;
+ virtual void getFluidNodeIndicesApplyBodyForce(uint *fluidNodeIndicesApplyBodyForce) const = 0;
+ virtual void getFluidNodeIndicesAllFeatures(uint *fluidNodeIndicesAllFeatures) const = 0;
+
};
#endif
diff --git a/src/gpu/GridGenerator/grid/GridBuilder/GridBuilder.h b/src/gpu/GridGenerator/grid/GridBuilder/GridBuilder.h
index 739aef59f76a33fa67d472a77ef258469f5e411c..065665d216e3cf7904530b94c8bb6480bb565c8a 100644
--- a/src/gpu/GridGenerator/grid/GridBuilder/GridBuilder.h
+++ b/src/gpu/GridGenerator/grid/GridBuilder/GridBuilder.h
@@ -54,6 +54,7 @@ class GridWrapper;
class Transformator;
class ArrowTransformator;
class PolyDataWriterWrapper;
+class TransientBCInputFileReader;
class BoundingBox;
class Grid;
@@ -113,6 +114,15 @@ public:
virtual void getPressureValues(real *rho, int *indices, int *neighborIndices, int level) const = 0;
virtual void getPressureQs(real *qs[27], int level) const = 0;
+ virtual uint getPrecursorSize(int level) const = 0;
+ virtual void getPrecursorValues(uint* neighbor0PP, uint* neighbor0PM, uint* neighbor0MP, uint* neighbor0MM,
+ real* weights0PP, real* weights0PM, real* weights0MP, real* weights0MM,
+ int* indices, std::vector<SPtr<TransientBCInputFileReader>>& reader,
+ int& numberOfPrecursorNodes, size_t& numberOfQuantities, uint& timeStepsBetweenReads,
+ real& velocityX, real& velocityY, real& velocityZ, int level) const = 0;
+
+ virtual void getPrecursorQs(real* qs[27], int level) const = 0;
+
virtual uint getGeometrySize(int level) const = 0;
virtual void getGeometryIndices(int *indices, int level) const = 0;
virtual void getGeometryQs(real *qs[27], int level) const = 0;
@@ -136,6 +146,21 @@ public:
virtual void getReceiveIndices(int *sendIndices, int direction, int level) = 0;
virtual void findFluidNodes(bool splitDomain) = 0;
+
+ virtual void addFluidNodeIndicesMacroVars(const std::vector<uint>& fluidNodeIndicesMacroVars, uint level) = 0;
+ virtual void addFluidNodeIndicesApplyBodyForce(const std::vector<uint>& fluidNodeIndicesApplyBodyForce, uint level) = 0;
+ virtual void addFluidNodeIndicesAllFeatures(const std::vector<uint>& fluidNodeIndicesAllFeatures, uint level) = 0;
+ virtual void sortFluidNodeIndicesMacroVars(uint level) = 0;
+ virtual void sortFluidNodeIndicesApplyBodyForce(uint level) = 0;
+ virtual void sortFluidNodeIndicesAllFeatures(uint level) = 0;
+ virtual uint getNumberOfFluidNodesMacroVars(uint level) const = 0;
+ virtual void getFluidNodeIndicesMacroVars(uint *fluidNodeIndicesMacroVars, int level) const = 0;
+ virtual uint getNumberOfFluidNodesApplyBodyForce(uint level) const = 0;
+ virtual void getFluidNodeIndicesApplyBodyForce(uint *fluidNodeIndicesApplyBodyForce, int level) const = 0;
+ virtual uint getNumberOfFluidNodesAllFeatures(uint level) const = 0;
+ virtual void getFluidNodeIndicesAllFeatures(uint *fluidNodeIndicesAllFeatures, int level) const = 0;
+
+
};
#endif
\ No newline at end of file
diff --git a/src/gpu/GridGenerator/grid/GridBuilder/LevelGridBuilder.cpp b/src/gpu/GridGenerator/grid/GridBuilder/LevelGridBuilder.cpp
index 083b9a51e0b151f49922df456e968c4b204e4af7..ff6f4913e4cd8a32c05272ef583f90a2cf226edc 100644
--- a/src/gpu/GridGenerator/grid/GridBuilder/LevelGridBuilder.cpp
+++ b/src/gpu/GridGenerator/grid/GridBuilder/LevelGridBuilder.cpp
@@ -52,6 +52,8 @@
#include "io/QLineWriter.h"
#include "io/SimulationFileWriter/SimulationFileWriter.h"
+#include "TransientBCSetter/TransientBCSetter.h"
+
#include "utilities/communication.h"
#include "utilities/transformator/ArrowTransformator.h"
@@ -117,14 +119,19 @@ void LevelGridBuilder::setSlipGeometryBoundaryCondition(real normalX, real norma
}
}
+//=======================================================================================
+//! \brief Set stress boundary concdition using iMEM
+//! \param samplingOffset number of grid points above boundary where velocity for wall model is sampled
+//! \param z0 roughness length [m]
+//! \param dx dx of level 0 [m]
+//!
void LevelGridBuilder::setStressBoundaryCondition( SideType sideType,
real nomalX, real normalY, real normalZ,
- uint samplingOffset, real z0)
+ uint samplingOffset, real z0, real dx)
{
for (uint level = 0; level < getNumberOfGridLevels(); level++)
{
- SPtr<StressBoundaryCondition> stressBoundaryCondition = StressBoundaryCondition::make(nomalX, normalY, normalZ, samplingOffset, z0);
-
+ SPtr<StressBoundaryCondition> stressBoundaryCondition = StressBoundaryCondition::make(nomalX, normalY, normalZ, samplingOffset, z0*pow(2.0f,level)/dx);
auto side = SideFactory::make(sideType);
stressBoundaryCondition->side = side;
@@ -243,6 +250,39 @@ void LevelGridBuilder::setNoSlipGeometryBoundaryCondition()
}
}
+void LevelGridBuilder::setPrecursorBoundaryCondition(SideType sideType, SPtr<FileCollection> fileCollection, int timeStepsBetweenReads,
+ real velocityX, real velocityY, real velocityZ, std::vector<uint> fileLevelToGridLevelMap)
+{
+ if(fileLevelToGridLevelMap.empty())
+ {
+ *logging::out << logging::Logger::INFO_INTERMEDIATE << "Mapping precursor file levels to the corresponding grid levels" << "\n";
+
+ for (uint level = 0; level < getNumberOfGridLevels(); level++)
+ fileLevelToGridLevelMap.push_back(level);
+ }
+ else
+ {
+ if(fileLevelToGridLevelMap.size()!=getNumberOfGridLevels())
+ throw std::runtime_error("In setPrecursorBoundaryCondition: fileLevelToGridLevelMap does not match with the number of levels");
+ *logging::out << logging::Logger::INFO_INTERMEDIATE << "Using user defined file to grid level mapping" << "\n";
+ }
+
+ for (uint level = 0; level < getNumberOfGridLevels(); level++)
+ {
+ auto reader = createReaderForCollection(fileCollection, fileLevelToGridLevelMap[level]);
+ SPtr<PrecursorBoundaryCondition> precursorBoundaryCondition = PrecursorBoundaryCondition::make( reader, timeStepsBetweenReads, velocityX, velocityY, velocityZ);
+
+ auto side = SideFactory::make(sideType);
+
+ precursorBoundaryCondition->side = side;
+ precursorBoundaryCondition->side->addIndices(grids, level, precursorBoundaryCondition);
+
+ boundaryConditions[level]->precursorBoundaryConditions.push_back(precursorBoundaryCondition);
+
+ *logging::out << logging::Logger::INFO_INTERMEDIATE << "Set Precursor BC on level " << level << " with " << (int)precursorBoundaryCondition->indices.size() << "\n";
+ }
+}
+
GRIDGENERATOR_EXPORT void LevelGridBuilder::setEnableFixRefinementIntoTheWall(bool enableFixRefinementIntoTheWall)
{
for (uint level = 0; level < this->grids.size(); level++)
@@ -594,6 +634,86 @@ void LevelGridBuilder::getPressureQs(real* qs[27], int level) const
}
}
+uint LevelGridBuilder::getPrecursorSize(int level) const
+{
+ uint size = 0;
+ for (auto boundaryCondition : boundaryConditions[level]->precursorBoundaryConditions)
+ {
+ size += uint(boundaryCondition->indices.size());
+ }
+ return size;
+}
+
+void LevelGridBuilder::getPrecursorValues( uint* neighbor0PP, uint* neighbor0PM, uint* neighbor0MP, uint* neighbor0MM,
+ real* weights0PP, real* weights0PM, real* weights0MP, real* weights0MM,
+ int* indices, std::vector<SPtr<TransientBCInputFileReader>>& reader,
+ int& numberOfPrecursorNodes, size_t& numberOfQuantities, uint& timeStepsBetweenReads,
+ real& velocityX, real& velocityY, real& velocityZ, int level) const
+{
+ int allIndicesCounter = 0;
+ int allNodesCounter = 0;
+ uint tmpTimeStepsBetweenReads = 0;
+ size_t tmpNumberOfQuantities = 0;
+
+ for (auto boundaryCondition : boundaryConditions[level]->precursorBoundaryConditions)
+ {
+ if( tmpTimeStepsBetweenReads == 0 )
+ tmpTimeStepsBetweenReads = boundaryCondition->timeStepsBetweenReads;
+ if( tmpTimeStepsBetweenReads != boundaryCondition->timeStepsBetweenReads )
+ throw std::runtime_error("All precursor boundary conditions must have the same timeStepsBetweenReads value");
+ auto BCreader = boundaryCondition->getReader();
+ BCreader->setWritingOffset(allIndicesCounter);
+ reader.push_back(BCreader);
+
+ std::vector<real> y, z;
+ real xTmp, yTmp, zTmp;
+ for(uint i = 0; i<boundaryCondition->indices.size(); i++)
+ {
+ indices[allIndicesCounter] = grids[level]->getSparseIndex(boundaryCondition->indices[i]) + 1;
+ grids[level]->transIndexToCoords(boundaryCondition->indices[i], xTmp, yTmp, zTmp);
+ y.push_back(yTmp);
+ z.push_back(zTmp);
+ allIndicesCounter++;
+ }
+ BCreader->fillArrays(y, z);
+ BCreader->getNeighbors(neighbor0PP, neighbor0PM, neighbor0MP, neighbor0MM);
+ BCreader->getWeights(weights0PP, weights0PM, weights0MP, weights0MM);
+ if(tmpNumberOfQuantities == 0)
+ tmpNumberOfQuantities = BCreader->getNumberOfQuantities();
+ if(tmpNumberOfQuantities != BCreader->getNumberOfQuantities())
+ throw std::runtime_error("All precursor files must have the same quantities.");
+ allNodesCounter += BCreader->getNPointsRead();
+ velocityX = boundaryCondition->getVelocityX();
+ velocityY = boundaryCondition->getVelocityY();
+ velocityZ = boundaryCondition->getVelocityZ();
+ }
+ numberOfPrecursorNodes = allNodesCounter;
+
+ if (tmpTimeStepsBetweenReads == 0)
+ throw std::runtime_error("timeStepsBetweenReads of precursor needs to be larger than 0.");
+ timeStepsBetweenReads = tmpTimeStepsBetweenReads;
+
+ if (tmpNumberOfQuantities == 0)
+ throw std::runtime_error("Number of quantities in precursor needs to be larger than 0.");
+ numberOfQuantities = tmpNumberOfQuantities;
+}
+
+void LevelGridBuilder::getPrecursorQs(real* qs[27], int level) const
+{
+ int allIndicesCounter = 0;
+ for (auto boundaryCondition : boundaryConditions[level]->precursorBoundaryConditions)
+ {
+ for ( uint index = 0; index < boundaryCondition->indices.size(); index++ )
+ {
+ for (int dir = 0; dir <= grids[level]->getEndDirection(); dir++)
+ {
+ qs[dir][allIndicesCounter] = boundaryCondition->qs[index][dir];
+ }
+ allIndicesCounter++;
+ }
+ }
+}
+
uint LevelGridBuilder::getGeometrySize(int level) const
{
if (boundaryConditions[level]->geometryBoundaryCondition)
@@ -674,4 +794,65 @@ void LevelGridBuilder::findFluidNodes(bool splitDomain)
for (uint i = 0; i < grids.size(); i++)
grids[i]->findFluidNodeIndices(splitDomain);
*logging::out << logging::Logger::INFO_HIGH << "Done with findFluidNodes()\n";
+}
+
+
+void LevelGridBuilder::addFluidNodeIndicesMacroVars(const std::vector<uint>& fluidNodeIndicesMacroVars, uint level)
+{
+ grids[level]->addFluidNodeIndicesMacroVars(fluidNodeIndicesMacroVars);
+}
+
+void LevelGridBuilder::addFluidNodeIndicesApplyBodyForce(const std::vector<uint>& fluidNodeIndicesApplyBodyForce, uint level)
+{
+ grids[level]->addFluidNodeIndicesApplyBodyForce(fluidNodeIndicesApplyBodyForce);
+}
+
+void LevelGridBuilder::addFluidNodeIndicesAllFeatures(const std::vector<uint>& fluidNodeIndicesAllFeatures, uint level)
+{
+ grids[level]->addFluidNodeIndicesAllFeatures(fluidNodeIndicesAllFeatures);
+}
+
+void LevelGridBuilder::sortFluidNodeIndicesMacroVars(uint level)
+{
+ grids[level]->sortFluidNodeIndicesMacroVars();
+}
+
+void LevelGridBuilder::sortFluidNodeIndicesApplyBodyForce(uint level)
+{
+ grids[level]->sortFluidNodeIndicesApplyBodyForce();
+}
+
+void LevelGridBuilder::sortFluidNodeIndicesAllFeatures(uint level)
+{
+ grids[level]->sortFluidNodeIndicesAllFeatures();
+}
+
+uint LevelGridBuilder::getNumberOfFluidNodesMacroVars(unsigned int level) const
+{
+ return grids[level]->getNumberOfFluidNodeIndicesMacroVars();
+}
+
+void LevelGridBuilder::getFluidNodeIndicesMacroVars(uint *fluidNodeIndicesMacroVars, const int level) const
+{
+ grids[level]->getFluidNodeIndicesMacroVars(fluidNodeIndicesMacroVars);
+}
+
+uint LevelGridBuilder::getNumberOfFluidNodesApplyBodyForce(unsigned int level) const
+{
+ return grids[level]->getNumberOfFluidNodeIndicesApplyBodyForce();
+}
+
+void LevelGridBuilder::getFluidNodeIndicesApplyBodyForce(uint *fluidNodeIndicesApplyBodyForce, const int level) const
+{
+ grids[level]->getFluidNodeIndicesApplyBodyForce(fluidNodeIndicesApplyBodyForce);
+}
+
+uint LevelGridBuilder::getNumberOfFluidNodesAllFeatures(unsigned int level) const
+{
+ return grids[level]->getNumberOfFluidNodeIndicesAllFeatures();
+}
+
+void LevelGridBuilder::getFluidNodeIndicesAllFeatures(uint *fluidNodeIndicesAllFeatures, const int level) const
+{
+ grids[level]->getFluidNodeIndicesAllFeatures(fluidNodeIndicesAllFeatures);
}
\ No newline at end of file
diff --git a/src/gpu/GridGenerator/grid/GridBuilder/LevelGridBuilder.h b/src/gpu/GridGenerator/grid/GridBuilder/LevelGridBuilder.h
index afb027fc1665ab874523bf39ec2a05518d28f7a1..56ae1e4fce6185591fba97f49ba504ced259aea5 100644
--- a/src/gpu/GridGenerator/grid/GridBuilder/LevelGridBuilder.h
+++ b/src/gpu/GridGenerator/grid/GridBuilder/LevelGridBuilder.h
@@ -38,6 +38,8 @@
#include <memory>
#include <array>
+#include <lbm/constants/NumericConstants.h>
+
#include "gpu/GridGenerator/global.h"
#include "gpu/GridGenerator/grid/GridBuilder/GridBuilder.h"
@@ -45,6 +47,8 @@
#include "gpu/GridGenerator/grid/GridInterface.h"
#include "gpu/GridGenerator/grid/NodeValues.h"
+using namespace vf::lbm::constant;
+
struct Vertex;
class Grid;
class Transformator;
@@ -58,9 +62,11 @@ class SlipBoundaryCondition;
class StressBoundaryCondition;
class PressureBoundaryCondition;
class GeometryBoundaryCondition;
+class PrecursorBoundaryCondition;
enum class SideType;
-
+class TransientBCInputFileReader;
+class FileCollection;
class LevelGridBuilder : public GridBuilder
{
@@ -75,11 +81,14 @@ public:
GRIDGENERATOR_EXPORT ~LevelGridBuilder() override;
GRIDGENERATOR_EXPORT void setSlipBoundaryCondition(SideType sideType, real nomalX, real normalY, real normalZ);
- GRIDGENERATOR_EXPORT void setStressBoundaryCondition(SideType sideType, real nomalX, real normalY, real normalZ, uint samplingOffset, real z0);
+ GRIDGENERATOR_EXPORT void setStressBoundaryCondition(SideType sideType, real nomalX, real normalY, real normalZ, uint samplingOffset, real z0, real dx);
GRIDGENERATOR_EXPORT void setVelocityBoundaryCondition(SideType sideType, real vx, real vy, real vz);
GRIDGENERATOR_EXPORT void setPressureBoundaryCondition(SideType sideType, real rho);
GRIDGENERATOR_EXPORT void setPeriodicBoundaryCondition(bool periodic_X, bool periodic_Y, bool periodic_Z);
GRIDGENERATOR_EXPORT void setNoSlipBoundaryCondition(SideType sideType);
+ GRIDGENERATOR_EXPORT void setPrecursorBoundaryCondition(SideType sideType, SPtr<FileCollection> fileCollection, int timeStepsBetweenReads,
+ real velocityX=c0o1, real velocityY=c0o1, real velocityZ=c0o1,
+ std::vector<uint> fileLevelToGridLevelMap = {});
GRIDGENERATOR_EXPORT void setEnableFixRefinementIntoTheWall(bool enableFixRefinementIntoTheWall);
@@ -121,6 +130,14 @@ public:
GRIDGENERATOR_EXPORT void getPressureValues(real* rho, int* indices, int* neighborIndices, int level) const override;
GRIDGENERATOR_EXPORT virtual void getPressureQs(real* qs[27], int level) const override;
+ GRIDGENERATOR_EXPORT uint getPrecursorSize(int level) const override;
+ GRIDGENERATOR_EXPORT void getPrecursorValues( uint* neighbor0PP, uint* neighbor0PM, uint* neighbor0MP, uint* neighbor0MM,
+ real* weights0PP, real* weights0PM, real* weights0MP, real* weights0MM,
+ int* indices, std::vector<SPtr<TransientBCInputFileReader>>& reader,
+ int& numberOfPrecursorNodes, size_t& numberOfQuantities, uint& timeStepsBetweenReads,
+ real& velocityX, real& velocityY, real& velocityZ, int level) const override;
+ GRIDGENERATOR_EXPORT virtual void getPrecursorQs(real* qs[27], int level) const override;
+
GRIDGENERATOR_EXPORT virtual void getGeometryQs(real *qs[27], int level) const override;
GRIDGENERATOR_EXPORT virtual uint getGeometrySize(int level) const override;
GRIDGENERATOR_EXPORT virtual void getGeometryIndices(int *indices, int level) const override;
@@ -149,6 +166,8 @@ protected:
std::vector<SPtr<VelocityBoundaryCondition>> noSlipBoundaryConditions;
+ std::vector<SPtr<PrecursorBoundaryCondition>> precursorBoundaryConditions;
+
SPtr<GeometryBoundaryCondition> geometryBoundaryCondition;
};
bool geometryHasValues = false;
@@ -194,6 +213,21 @@ public:
// needed for CUDA Streams MultiGPU (Communication Hiding)
void findFluidNodes(bool splitDomain) override;
+
+ void addFluidNodeIndicesMacroVars(const std::vector<uint>& fluidNodeIndicesMacroVars, uint level) override;
+ void addFluidNodeIndicesApplyBodyForce(const std::vector<uint>& fluidNodeIndicesApplyBodyForce, uint level) override;
+ void addFluidNodeIndicesAllFeatures(const std::vector<uint>& fluidNodeIndicesAllFeatures, uint level) override;
+
+ void sortFluidNodeIndicesMacroVars(uint level) override;
+ void sortFluidNodeIndicesApplyBodyForce(uint level) override;
+ void sortFluidNodeIndicesAllFeatures(uint level) override;
+
+ uint getNumberOfFluidNodesMacroVars(unsigned int level) const override;
+ void getFluidNodeIndicesMacroVars(uint *fluidNodeIndicesMacroVars, const int level) const override;
+ uint getNumberOfFluidNodesApplyBodyForce(unsigned int level) const override;
+ void getFluidNodeIndicesApplyBodyForce(uint *fluidNodeIndicesApplyBodyForce, const int level) const override;
+ uint getNumberOfFluidNodesAllFeatures(unsigned int level) const override;
+ void getFluidNodeIndicesAllFeatures(uint *fluidNodeIndicesAllFeatures, const int level) const override;
};
#endif
diff --git a/src/gpu/GridGenerator/grid/GridImp.cpp b/src/gpu/GridGenerator/grid/GridImp.cpp
index 31bbf3ddc87184846fcb01a3e6631358b6a6f864..24dd169e70288b9ae18a29c23dc0c74b14246e3f 100644
--- a/src/gpu/GridGenerator/grid/GridImp.cpp
+++ b/src/gpu/GridGenerator/grid/GridImp.cpp
@@ -1154,7 +1154,7 @@ void GridImp::limitToSubDomain(SPtr<BoundingBox> subDomainBox, LbmOrGks lbmOrGks
// one layer for receive nodes and one for stoppers
if( lbmOrGks == LBM )
tmpSubDomainBox.extend(this->delta);
-
+
if (!tmpSubDomainBox.isInside(x, y, z)
&& ( this->getFieldEntry(index) == FLUID ||
this->getFieldEntry(index) == FLUID_CFC ||
@@ -1162,7 +1162,7 @@ void GridImp::limitToSubDomain(SPtr<BoundingBox> subDomainBox, LbmOrGks lbmOrGks
this->getFieldEntry(index) == FLUID_FCC ||
this->getFieldEntry(index) == FLUID_FCF ||
this->getFieldEntry(index) == BC_SOLID ) )
- {
+ {
this->setFieldEntry(index, STOPPER_OUT_OF_GRID_BOUNDARY);
}
}
@@ -1648,10 +1648,9 @@ bool GridImp::checkIfAtLeastOneValidQ(const uint index, const Vertex & point, co
void GridImp::findCommunicationIndices(int direction, SPtr<BoundingBox> subDomainBox, LbmOrGks lbmOrGks)
{
for( uint index = 0; index < this->size; index++ ){
-
real x, y, z;
this->transIndexToCoords(index, x, y, z);
-
+
if( this->getFieldEntry(index) == INVALID_OUT_OF_GRID ||
this->getFieldEntry(index) == INVALID_SOLID ||
this->getFieldEntry(index) == INVALID_COARSE_UNDER_FINE ||
@@ -1660,7 +1659,6 @@ void GridImp::findCommunicationIndices(int direction, SPtr<BoundingBox> subDomai
if( lbmOrGks == LBM && this->getFieldEntry(index) == STOPPER_OUT_OF_GRID_BOUNDARY ) continue;
if( lbmOrGks == LBM && this->getFieldEntry(index) == STOPPER_SOLID ) continue;
-
if( direction == CommunicationDirections::MX ) findCommunicationIndex( index, x, subDomainBox->minX, direction);
if( direction == CommunicationDirections::PX ) findCommunicationIndex( index, x, subDomainBox->maxX, direction);
if( direction == CommunicationDirections::MY ) findCommunicationIndex( index, y, subDomainBox->minY, direction);
@@ -1674,14 +1672,11 @@ void GridImp::findCommunicationIndex( uint index, real coordinate, real limit, i
// negative direction get a negative sign
real s = ( direction % 2 == 0 ) ? ( -1.0 ) : ( 1.0 );
-
- if (std::abs(coordinate - (limit + s * 0.5 * this->delta)) < 0.1 * this->delta) {
+ if (std::abs(coordinate - (limit + s * 0.5 * this->delta)) < 0.1 * this->delta)
this->communicationIndices[direction].receiveIndices.push_back(index);
- }
- if (std::abs(coordinate - (limit - s * 0.5 * this->delta)) < 0.1 * this->delta) {
+ if (std::abs(coordinate - (limit - s * 0.5 * this->delta)) < 0.1 * this->delta)
this->communicationIndices[direction].sendIndices.push_back(index);
- }
}
bool GridImp::isSendNode(int index) const
@@ -2080,6 +2075,124 @@ void GridImp::getFluidNodeIndicesBorder(uint *fluidNodeIndicesBorder) const
fluidNodeIndicesBorder[nodeNumber] = this->fluidNodeIndicesBorder[nodeNumber];
}
+void GridImp::addFluidNodeIndicesMacroVars(std::vector<uint> _fluidNodeIndicesMacroVars)
+{
+ size_t newSize = this->fluidNodeIndicesMacroVars.size()+_fluidNodeIndicesMacroVars.size();
+ this->fluidNodeIndicesMacroVars.reserve(newSize);
+ std::copy(_fluidNodeIndicesMacroVars.begin(), _fluidNodeIndicesMacroVars.end(), std::back_inserter(this->fluidNodeIndicesMacroVars));
+}
+
+void GridImp::addFluidNodeIndicesApplyBodyForce(std::vector<uint> _fluidNodeIndicesApplyBodyForce)
+{
+
+ size_t newSize = this->fluidNodeIndicesApplyBodyForce.size()+_fluidNodeIndicesApplyBodyForce.size();
+ this->fluidNodeIndicesApplyBodyForce.reserve(newSize);
+ std::copy(_fluidNodeIndicesApplyBodyForce.begin(), _fluidNodeIndicesApplyBodyForce.end(), std::back_inserter(this->fluidNodeIndicesApplyBodyForce));
+}
+
+void GridImp::addFluidNodeIndicesAllFeatures(std::vector<uint> _fluidNodeIndicesAllFeatures)
+{
+
+ size_t newSize = this->fluidNodeIndicesAllFeatures.size()+_fluidNodeIndicesAllFeatures.size();
+ this->fluidNodeIndicesAllFeatures.reserve(newSize);
+ std::copy(_fluidNodeIndicesAllFeatures.begin(), _fluidNodeIndicesAllFeatures.end(), std::back_inserter(this->fluidNodeIndicesAllFeatures));
+}
+
+void GridImp::sortFluidNodeIndicesMacroVars()
+{
+ if(this->fluidNodeIndicesMacroVars.size()>0)
+ {
+ sort(this->fluidNodeIndicesMacroVars.begin(), this->fluidNodeIndicesMacroVars.end());
+ // Remove duplicates
+ this->fluidNodeIndicesMacroVars.erase( unique( this->fluidNodeIndicesMacroVars.begin(), this->fluidNodeIndicesMacroVars.end() ), this->fluidNodeIndicesMacroVars.end() );
+
+ // Remove indices of fluidNodeIndicesAllFeatures from fluidNodeIndicesMacroVars
+ if(this->fluidNodeIndicesAllFeatures.size()>0)
+ {
+ this->fluidNodeIndicesMacroVars.erase( std::remove_if( this->fluidNodeIndicesMacroVars.begin(), this->fluidNodeIndicesMacroVars.end(),
+ [&](auto x){return binary_search(fluidNodeIndicesAllFeatures.begin(),fluidNodeIndicesAllFeatures.end(),x);} ),
+ this->fluidNodeIndicesMacroVars.end()
+ );
+ }
+
+ // Remove indices of fluidNodeIndicesMacroVars from fluidNodeIndices
+ this->fluidNodeIndices.erase( std::remove_if( this->fluidNodeIndices.begin(), this->fluidNodeIndices.end(),
+ [&](auto x){return binary_search(fluidNodeIndicesMacroVars.begin(),fluidNodeIndicesMacroVars.end(),x);} ),
+ this->fluidNodeIndices.end()
+ );
+ }
+}
+
+void GridImp::sortFluidNodeIndicesApplyBodyForce()
+{
+ if(this->fluidNodeIndicesApplyBodyForce.size()>0)
+ {
+ sort(this->fluidNodeIndicesApplyBodyForce.begin(), this->fluidNodeIndicesApplyBodyForce.end());
+ // Remove duplicates
+ this->fluidNodeIndicesApplyBodyForce.erase( unique( this->fluidNodeIndicesApplyBodyForce.begin(), this->fluidNodeIndicesApplyBodyForce.end() ), this->fluidNodeIndicesApplyBodyForce.end() );
+
+ // Remove indices of fluidNodeIndicesAllFeatures from fluidNodeIndicesMacroVars
+ if(this->fluidNodeIndicesAllFeatures.size()>0)
+ {
+ this->fluidNodeIndicesApplyBodyForce.erase( std::remove_if( this->fluidNodeIndicesApplyBodyForce.begin(), this->fluidNodeIndicesApplyBodyForce.end(),
+ [&](auto x){return binary_search(fluidNodeIndicesAllFeatures.begin(),fluidNodeIndicesAllFeatures.end(),x);} ),
+ this->fluidNodeIndicesApplyBodyForce.end()
+ );
+ }
+
+ // Remove indices of fluidNodeIndicesMacroVars from fluidNodeIndices
+ this->fluidNodeIndices.erase( std::remove_if( this->fluidNodeIndices.begin(), this->fluidNodeIndices.end(),
+ [&](auto x){return binary_search(fluidNodeIndicesApplyBodyForce.begin(),fluidNodeIndicesApplyBodyForce.end(),x);} ),
+ this->fluidNodeIndices.end()
+ );
+ }
+}
+
+void GridImp::sortFluidNodeIndicesAllFeatures()
+{
+ if(this->fluidNodeIndicesAllFeatures.size()>0)
+ {
+ sort(this->fluidNodeIndicesAllFeatures.begin(), this->fluidNodeIndicesAllFeatures.end());
+ // Remove duplicates
+ this->fluidNodeIndicesAllFeatures.erase( unique( this->fluidNodeIndicesAllFeatures.begin(), this->fluidNodeIndicesAllFeatures.end() ), this->fluidNodeIndicesAllFeatures.end() );
+ // Remove indices of fluidNodeIndicesMacroVars from fluidNodeIndices
+ this->fluidNodeIndices.erase( std::remove_if( this->fluidNodeIndices.begin(), this->fluidNodeIndices.end(),
+ [&](auto x){return binary_search(fluidNodeIndicesAllFeatures.begin(),fluidNodeIndicesAllFeatures.end(),x);} ),
+ this->fluidNodeIndices.end()
+ );
+ }
+}
+
+uint GridImp::getNumberOfFluidNodeIndicesMacroVars() const {
+ return (uint)this->fluidNodeIndicesMacroVars.size();
+}
+
+uint GridImp::getNumberOfFluidNodeIndicesApplyBodyForce() const {
+ return (uint)this->fluidNodeIndicesApplyBodyForce.size();
+}
+
+uint GridImp::getNumberOfFluidNodeIndicesAllFeatures() const {
+ return (uint)this->fluidNodeIndicesAllFeatures.size();
+}
+
+void GridImp::getFluidNodeIndicesMacroVars(uint *_fluidNodeIndicesMacroVars) const
+{
+ std::copy(fluidNodeIndicesMacroVars.begin(), fluidNodeIndicesMacroVars.end(), _fluidNodeIndicesMacroVars);
+}
+void GridImp::getFluidNodeIndicesApplyBodyForce(uint *_fluidNodeIndicesApplyBodyForce) const
+{
+ std::copy(fluidNodeIndicesApplyBodyForce.begin(), fluidNodeIndicesApplyBodyForce.end(), _fluidNodeIndicesApplyBodyForce);
+}
+void GridImp::getFluidNodeIndicesAllFeatures(uint *_fluidNodeIndicesAllFeatures) const
+{
+ std::copy(fluidNodeIndicesAllFeatures.begin(), fluidNodeIndicesAllFeatures.end(), _fluidNodeIndicesAllFeatures);
+}
+
+
+
+
+
+
void GridImp::print() const
{
printf("min: (%2.4f, %2.4f, %2.4f), max: (%2.4f, %2.4f, %2.4f), size: %d, delta: %2.4f\n", startX, startY, startZ,
diff --git a/src/gpu/GridGenerator/grid/GridImp.h b/src/gpu/GridGenerator/grid/GridImp.h
index edb5ca916bf68dcf992ea214dcddb2dc43810352..8283bf569e266b84f020334a306d93756b01c394 100644
--- a/src/gpu/GridGenerator/grid/GridImp.h
+++ b/src/gpu/GridGenerator/grid/GridImp.h
@@ -115,8 +115,11 @@ private:
int *sparseIndices;
- std::vector<uint> fluidNodeIndices;
- std::vector<uint> fluidNodeIndicesBorder;
+ std::vector<uint> fluidNodeIndices; // run on CollisionTemplate::Default
+ std::vector<uint> fluidNodeIndicesBorder; // run on subdomain border nodes (CollisionTemplate::SubDomainBorder)
+ std::vector<uint> fluidNodeIndicesMacroVars; // run on CollisionTemplate::MacroVars
+ std::vector<uint> fluidNodeIndicesApplyBodyForce; // run on CollisionTemplate::ApplyBodyForce
+ std::vector<uint> fluidNodeIndicesAllFeatures; // run on CollisionTemplate::AllFeatures
uint *qIndices; //maps from matrix index to qIndex
real *qValues;
@@ -364,6 +367,19 @@ public:
uint getNumberOfFluidNodesBorder() const override;
void getFluidNodeIndicesBorder(uint *fluidNodeIndicesBorder) const override;
+ void addFluidNodeIndicesMacroVars(std::vector<uint> _fluidNodeIndicesMacroVars) override;
+ void addFluidNodeIndicesApplyBodyForce(std::vector<uint> _fluidNodeIndicesApplyBodyForce) override;
+ void addFluidNodeIndicesAllFeatures(std::vector<uint> _fluidNodeIndicesAllFeatures) override;
+ void sortFluidNodeIndicesMacroVars() override;
+ void sortFluidNodeIndicesApplyBodyForce() override;
+ void sortFluidNodeIndicesAllFeatures() override;
+
+ uint getNumberOfFluidNodeIndicesMacroVars() const override;
+ uint getNumberOfFluidNodeIndicesApplyBodyForce() const override;
+ uint getNumberOfFluidNodeIndicesAllFeatures() const override;
+ void getFluidNodeIndicesMacroVars(uint *fluidNodeIndicesMacroVars) const override;
+ void getFluidNodeIndicesApplyBodyForce(uint *fluidNodeIndicesApplyBodyForce) const override;
+ void getFluidNodeIndicesAllFeatures(uint *fluidNodeIndicesAllFeatures) const override;
public:
struct CommunicationIndices {
diff --git a/src/gpu/VirtualFluids_GPU/Calculation/CollisisionStrategy.cpp b/src/gpu/VirtualFluids_GPU/Calculation/CollisisionStrategy.cpp
index 4a14d19c10936f84379f332ef24f081f0ebb0cb7..49543f37df7fb54290f4ab6c09edb8d10c0b67be 100644
--- a/src/gpu/VirtualFluids_GPU/Calculation/CollisisionStrategy.cpp
+++ b/src/gpu/VirtualFluids_GPU/Calculation/CollisisionStrategy.cpp
@@ -39,8 +39,14 @@ void CollisionAndExchange_noStreams_indexKernel::operator()(UpdateGrid27 *update
//!
//! 1. run collision
//!
- updateGrid->collisionUsingIndices(level, t, para->getParD(level)->fluidNodeIndices,
- para->getParD(level)->numberOfFluidNodes, -1);
+ for( CollisionTemplate tag: para->getParH(level)->allocatedBulkFluidNodeTags )
+ {
+ updateGrid->collisionUsingIndices( level, t,
+ para->getParD(level)->taggedFluidNodeIndices[tag],
+ para->getParD(level)->numberOfTaggedFluidNodes[tag],
+ tag,
+ CudaStreamIndex::Legacy);
+ }
//! 2. exchange information between GPUs
updateGrid->exchangeMultiGPU_noStreams_withPrepare(level, false);
@@ -61,28 +67,35 @@ void CollisionAndExchange_noStreams_oldKernel::operator()(UpdateGrid27 *updateGr
void CollisionAndExchange_streams::operator()(UpdateGrid27 *updateGrid, Parameter *para, int level, unsigned int t)
{
- int borderStreamIndex = para->getStreamManager()->getBorderStreamIndex();
- int bulkStreamIndex = para->getStreamManager()->getBulkStreamIndex();
-
//! \details steps:
//!
- //! 1. run collision for nodes which are at the border of the gpus/processes
- //!
- updateGrid->collisionUsingIndices(level, t, para->getParD(level)->fluidNodeIndicesBorder,
- para->getParD(level)->numberOfFluidNodesBorder, borderStreamIndex);
+ //! 1. run collision for nodes which are at the border of the gpus/processes, running with WriteMacroVars in case probes sample on these nodes
+ //!
+ updateGrid->collisionUsingIndices( level, t,
+ para->getParD(level)->taggedFluidNodeIndices[CollisionTemplate::SubDomainBorder],
+ para->getParD(level)->numberOfTaggedFluidNodes[CollisionTemplate::SubDomainBorder],
+ CollisionTemplate::WriteMacroVars,
+ CudaStreamIndex::SubDomainBorder);
//! 2. prepare the exchange between gpus (collect the send nodes for communication in a buffer on the gpu) and trigger bulk kernel execution when finished
//!
- updateGrid->prepareExchangeMultiGPU(level, borderStreamIndex);
+ updateGrid->prepareExchangeMultiGPU(level, CudaStreamIndex::SubDomainBorder);
if (para->getUseStreams())
- para->getStreamManager()->triggerStartBulkKernel(borderStreamIndex);
-
- //! 3. launch the collision kernel for bulk nodes
- //!
- para->getStreamManager()->waitOnStartBulkKernelEvent(bulkStreamIndex);
- updateGrid->collisionUsingIndices(level, t, para->getParD(level)->fluidNodeIndices,
- para->getParD(level)->numberOfFluidNodes, bulkStreamIndex);
-
+ para->getStreamManager()->triggerStartBulkKernel(CudaStreamIndex::SubDomainBorder);
+
+ //! 3. launch the collision kernel for bulk nodes. This includes nodes with \param tag Default, WriteMacroVars, ApplyBodyForce,
+ //! or AllFeatures. All assigned tags are listed in \param allocatedBulkFluidNodeTags during initialization in Simulation::init
+
+ para->getStreamManager()->waitOnStartBulkKernelEvent(CudaStreamIndex::Bulk);
+
+ for( CollisionTemplate tag: para->getParH(level)->allocatedBulkFluidNodeTags )
+ {
+ updateGrid->collisionUsingIndices( level, t,
+ para->getParD(level)->taggedFluidNodeIndices[tag],
+ para->getParD(level)->numberOfTaggedFluidNodes[tag],
+ tag,
+ CudaStreamIndex::Bulk);
+ }
//! 4. exchange information between GPUs
- updateGrid->exchangeMultiGPU(level, borderStreamIndex);
+ updateGrid->exchangeMultiGPU(level, CudaStreamIndex::SubDomainBorder);
}
diff --git a/src/gpu/VirtualFluids_GPU/Calculation/RefinementStrategy.cpp b/src/gpu/VirtualFluids_GPU/Calculation/RefinementStrategy.cpp
index cd74216e1fbe7b718c72046ace4b7d2e7cf451fe..b8ca4e9c2020e17cd0192267ac5d931b510afc3a 100644
--- a/src/gpu/VirtualFluids_GPU/Calculation/RefinementStrategy.cpp
+++ b/src/gpu/VirtualFluids_GPU/Calculation/RefinementStrategy.cpp
@@ -38,67 +38,62 @@ void NoRefinement::operator()(UpdateGrid27 *updateGrid, Parameter *para, int lev
void RefinementAndExchange_streams_exchangeInterface::operator()(UpdateGrid27 *updateGrid, Parameter *para, int level)
{
- int borderStreamIndex = para->getStreamManager()->getBorderStreamIndex();
- int bulkStreamIndex = para->getStreamManager()->getBulkStreamIndex();
-
//! \details steps:
//!
//! 1. Interpolation fine to coarse for nodes which are at the border of the gpus/processes
//!
- updateGrid->fineToCoarse(level, ¶->getParD(level)->intFCBorder, para->getParD(level)->offFC, borderStreamIndex);
+ updateGrid->fineToCoarse(level, ¶->getParD(level)->intFCBorder, para->getParD(level)->offFC, CudaStreamIndex::SubDomainBorder);
//! 2. prepare the exchange between gpus (collect the send nodes for communication in a buffer on the gpu) and trigger bulk kernel execution when finished
//!
- updateGrid->prepareExchangeMultiGPUAfterFtoC(level, borderStreamIndex);
+ updateGrid->prepareExchangeMultiGPUAfterFtoC(level, CudaStreamIndex::SubDomainBorder);
if (para->getUseStreams())
- para->getStreamManager()->triggerStartBulkKernel(borderStreamIndex);
+ para->getStreamManager()->triggerStartBulkKernel(CudaStreamIndex::SubDomainBorder);
//! 3. launch the bulk kernels for both interpolation processes (fine to coarse and coarse to fine)
//!
- para->getStreamManager()->waitOnStartBulkKernelEvent(bulkStreamIndex);
- updateGrid->fineToCoarse(level, ¶->getParD(level)->intFCBulk, para->getParD(level)->offFCBulk, bulkStreamIndex);
- updateGrid->coarseToFine(level, ¶->getParD(level)->intCFBulk, para->getParD(level)->offCFBulk, bulkStreamIndex);
+ para->getStreamManager()->waitOnStartBulkKernelEvent(CudaStreamIndex::Bulk);
+ updateGrid->fineToCoarse(level, ¶->getParD(level)->intFCBulk, para->getParD(level)->offFCBulk, CudaStreamIndex::SubDomainBorder);
+ updateGrid->coarseToFine(level, ¶->getParD(level)->intCFBulk, para->getParD(level)->offCFBulk, CudaStreamIndex::SubDomainBorder);
//! 4. exchange information between GPUs (only nodes which are part of the interpolation)
//!
- updateGrid->exchangeMultiGPUAfterFtoC(level, borderStreamIndex);
+ updateGrid->exchangeMultiGPUAfterFtoC(level, CudaStreamIndex::SubDomainBorder);
// 5. interpolation fine to coarse for nodes which are at the border of the gpus/processes
//!
- updateGrid->coarseToFine(level, ¶->getParD(level)->intCFBorder, para->getParD(level)->offCF, borderStreamIndex);
+ updateGrid->coarseToFine(level, ¶->getParD(level)->intCFBorder, para->getParD(level)->offCF, CudaStreamIndex::SubDomainBorder);
cudaDeviceSynchronize();
}
-void RefinementAndExchange_streams_exchangeAllNodes::operator()(UpdateGrid27 *updateGrid, Parameter *para, int level){
- int borderStreamIndex = para->getStreamManager()->getBorderStreamIndex();
- int bulkStreamIndex = para->getStreamManager()->getBulkStreamIndex();
-
+void RefinementAndExchange_streams_exchangeAllNodes::operator()(UpdateGrid27 *updateGrid, Parameter *para, int level)
+{
//! \details steps:
//!
//! 1. interpolation fine to coarse for nodes which are at the border of the gpus/processes
//!
- updateGrid->fineToCoarse(level, ¶->getParD(level)->intFCBorder, para->getParD(level)->offFC, borderStreamIndex);
+ updateGrid->fineToCoarse(level, ¶->getParD(level)->intFCBorder, para->getParD(level)->offFC, CudaStreamIndex::SubDomainBorder);
//! 2. prepare the exchange between gpus (collect the send nodes for communication in a buffer on the gpu) and trigger bulk kernel execution when finished
//!
- updateGrid->prepareExchangeMultiGPU(level, borderStreamIndex);
+ updateGrid->prepareExchangeMultiGPU(level, CudaStreamIndex::SubDomainBorder);
if (para->getUseStreams())
- para->getStreamManager()->triggerStartBulkKernel(borderStreamIndex);
+ para->getStreamManager()->triggerStartBulkKernel(CudaStreamIndex::SubDomainBorder);
//! 3. launch the bulk kernels for both interpolation processes (fine to coarse and coarse to fine)
//!
- para->getStreamManager()->waitOnStartBulkKernelEvent(bulkStreamIndex);
- updateGrid->fineToCoarse(level, ¶->getParD(level)->intFCBulk, para->getParD(level)->offFCBulk, bulkStreamIndex);
- updateGrid->coarseToFine(level, ¶->getParD(level)->intCFBulk, para->getParD(level)->offCFBulk, bulkStreamIndex);
+ para->getStreamManager()->waitOnStartBulkKernelEvent(CudaStreamIndex::Bulk);
+ updateGrid->fineToCoarse(level, ¶->getParD(level)->intFCBulk, para->getParD(level)->offFCBulk, CudaStreamIndex::SubDomainBorder);
+ updateGrid->coarseToFine(level, ¶->getParD(level)->intCFBulk, para->getParD(level)->offCFBulk, CudaStreamIndex::SubDomainBorder);
//! 4. exchange information between GPUs (all nodes)
//!
- updateGrid->exchangeMultiGPU(level, borderStreamIndex);
+ updateGrid->exchangeMultiGPU(level, CudaStreamIndex::SubDomainBorder);
// 5. interpolation fine to coarse for nodes which are at the border of the gpus/processes
//!
- updateGrid->coarseToFine(level, ¶->getParD(level)->intCFBorder, para->getParD(level)->offCF, borderStreamIndex);
+ updateGrid->coarseToFine(level, ¶->getParD(level)->intCFBorder, para->getParD(level)->offCF, CudaStreamIndex::SubDomainBorder);
cudaDeviceSynchronize();
}
@@ -109,14 +104,14 @@ void RefinementAndExchange_noStreams_exchangeInterface::operator()(UpdateGrid27
//!
//! 1. interpolation fine to coarse
//!
- updateGrid->fineToCoarse(level, ¶->getParD(level)->intFC, para->getParD(level)->offFC, -1);
+ updateGrid->fineToCoarse(level, ¶->getParD(level)->intFC, para->getParD(level)->offFC, CudaStreamIndex::Legacy);
//! 2. exchange information between GPUs (only nodes which are part of the interpolation)
//!
updateGrid->exchangeMultiGPU_noStreams_withPrepare(level, true);
//! 3. interpolation coarse to fine
- updateGrid->coarseToFine(level, ¶->getParD(level)->intCF, para->getParD(level)->offCF, -1);
+ updateGrid->coarseToFine(level, ¶->getParD(level)->intCF, para->getParD(level)->offCF, CudaStreamIndex::Legacy);
}
void RefinementAndExchange_noStreams_exchangeAllNodes::operator()(UpdateGrid27 *updateGrid, Parameter *para, int level)
@@ -125,14 +120,14 @@ void RefinementAndExchange_noStreams_exchangeAllNodes::operator()(UpdateGrid27 *
//!
//! 1. interpolation fine to coarse
//!
- updateGrid->fineToCoarse(level, ¶->getParD(level)->intFC, para->getParD(level)->offFC, -1);
+ updateGrid->fineToCoarse(level, ¶->getParD(level)->intFC, para->getParD(level)->offFC, CudaStreamIndex::Legacy);
//! 2. exchange information between GPUs (all nodes)
//!
updateGrid->exchangeMultiGPU_noStreams_withPrepare(level, false);
//! 3. interpolation coarse to fine
- updateGrid->coarseToFine(level, ¶->getParD(level)->intCF, para->getParD(level)->offCF, -1);
+ updateGrid->coarseToFine(level, ¶->getParD(level)->intCF, para->getParD(level)->offCF, CudaStreamIndex::Legacy);
}
void Refinement_noExchange::operator()(UpdateGrid27 *updateGrid, Parameter *para, int level)
@@ -141,7 +136,7 @@ void Refinement_noExchange::operator()(UpdateGrid27 *updateGrid, Parameter *para
//!
//! 1. interpolation fine to coarse
//!
- updateGrid->fineToCoarse(level, ¶->getParD(level)->intFC, para->getParD(level)->offFC, -1);
+ updateGrid->fineToCoarse(level, ¶->getParD(level)->intFC, para->getParD(level)->offFC, CudaStreamIndex::Legacy);
//! 2. interpolation coarse to fine
- updateGrid->coarseToFine(level, ¶->getParD(level)->intCF, para->getParD(level)->offCF, -1);
+ updateGrid->coarseToFine(level, ¶->getParD(level)->intCF, para->getParD(level)->offCF, CudaStreamIndex::Legacy);
}
diff --git a/src/gpu/VirtualFluids_GPU/Calculation/UpdateGrid27.cpp b/src/gpu/VirtualFluids_GPU/Calculation/UpdateGrid27.cpp
index 296ab819c5538a6b6d6a6827b5c28cbc475af838..4136614dfbfc9e0d2fc1bf7f4b01624f94eabb6f 100644
--- a/src/gpu/VirtualFluids_GPU/Calculation/UpdateGrid27.cpp
+++ b/src/gpu/VirtualFluids_GPU/Calculation/UpdateGrid27.cpp
@@ -22,13 +22,17 @@ void UpdateGrid27::updateGrid(int level, unsigned int t)
updateGrid(level + 1, t);
}
+ //////////////////////////////////////////////////////////////////////////
+
+ interactWithProbes(level, t);
+
//////////////////////////////////////////////////////////////////////////
collision(this, para.get(), level, t);
//////////////////////////////////////////////////////////////////////////
- postCollisionBC(level);
+ postCollisionBC(level, t);
//////////////////////////////////////////////////////////////////////////
@@ -47,13 +51,14 @@ void UpdateGrid27::updateGrid(int level, unsigned int t)
//////////////////////////////////////////////////////////////////////////
if( level != para->getFine() )
- {
+ {
refinement(this, para.get(), level);
}
+ //////////////////////////////////////////////////////////////////////////
+
interactWithActuators(level, t);
- interactWithProbes(level, t);
}
void UpdateGrid27::collisionAllNodes(int level, unsigned int t)
@@ -71,15 +76,16 @@ void UpdateGrid27::collisionAllNodes(int level, unsigned int t)
collisionAdvectionDiffusion(level);
}
-void UpdateGrid27::collisionUsingIndices(int level, unsigned int t, uint *fluidNodeIndices, uint numberOfFluidNodes, int stream)
+void UpdateGrid27::collisionUsingIndices(int level, unsigned int t, uint *taggedFluidNodeIndices, uint numberOfTaggedFluidNodes, CollisionTemplate collisionTemplate, CudaStreamIndex stream)
{
- if (fluidNodeIndices != nullptr && numberOfFluidNodes != 0)
- kernels.at(level)->runOnIndices(fluidNodeIndices, numberOfFluidNodes, stream);
+ if (taggedFluidNodeIndices != nullptr && numberOfTaggedFluidNodes != 0)
+ kernels.at(level)->runOnIndices(taggedFluidNodeIndices, numberOfTaggedFluidNodes, collisionTemplate, stream);
else
- std::cout << "In collision: fluidNodeIndices or numberOfFluidNodes not definded"
+ std::cout << "In collision: fluidNodeIndices or numberOfFluidNodes not defined"
<< std::endl;
//////////////////////////////////////////////////////////////////////////
+ //! \todo: AD collision and porousMedia should be called separately, not in collisionUsingIndices
if (para->getSimulatePorousMedia())
collisionPorousMedia(level);
@@ -118,21 +124,21 @@ void UpdateGrid27::collisionAdvectionDiffusion(int level)
this->adKernelManager->runADcollisionKernel(level);
}
-void UpdateGrid27::prepareExchangeMultiGPU(int level, int streamIndex)
+void UpdateGrid27::prepareExchangeMultiGPU(int level, CudaStreamIndex streamIndex)
{
prepareExchangeCollDataXGPU27AllNodes(para.get(), level, streamIndex);
prepareExchangeCollDataYGPU27AllNodes(para.get(), level, streamIndex);
prepareExchangeCollDataZGPU27AllNodes(para.get(), level, streamIndex);
}
-void UpdateGrid27::prepareExchangeMultiGPUAfterFtoC(int level, int streamIndex)
+void UpdateGrid27::prepareExchangeMultiGPUAfterFtoC(int level, CudaStreamIndex streamIndex)
{
prepareExchangeCollDataXGPU27AfterFtoC(para.get(), level, streamIndex);
prepareExchangeCollDataYGPU27AfterFtoC(para.get(), level, streamIndex);
prepareExchangeCollDataZGPU27AfterFtoC(para.get(), level, streamIndex);
}
-void UpdateGrid27::exchangeMultiGPU(int level, int streamIndex)
+void UpdateGrid27::exchangeMultiGPU(int level, CudaStreamIndex streamIndex)
{
//////////////////////////////////////////////////////////////////////////
// 3D domain decomposition
@@ -168,30 +174,30 @@ void UpdateGrid27::exchangeMultiGPU_noStreams_withPrepare(int level, bool useRed
// 3D domain decomposition
if (useReducedComm) {
// X
- prepareExchangeCollDataXGPU27AfterFtoC(para.get(), level, -1);
- exchangeCollDataXGPU27AfterFtoC(para.get(), comm, cudaMemoryManager.get(), level, -1);
- scatterNodesFromRecvBufferXGPU27AfterFtoC(para.get(), level, -1);
+ prepareExchangeCollDataXGPU27AfterFtoC(para.get(), level, CudaStreamIndex::Legacy);
+ exchangeCollDataXGPU27AfterFtoC(para.get(), comm, cudaMemoryManager.get(), level, CudaStreamIndex::Legacy);
+ scatterNodesFromRecvBufferXGPU27AfterFtoC(para.get(), level, CudaStreamIndex::Legacy);
// Y
- prepareExchangeCollDataYGPU27AfterFtoC(para.get(), level, -1);
- exchangeCollDataYGPU27AfterFtoC(para.get(), comm, cudaMemoryManager.get(), level, -1);
- scatterNodesFromRecvBufferYGPU27AfterFtoC(para.get(), level, -1);
+ prepareExchangeCollDataYGPU27AfterFtoC(para.get(), level, CudaStreamIndex::Legacy);
+ exchangeCollDataYGPU27AfterFtoC(para.get(), comm, cudaMemoryManager.get(), level, CudaStreamIndex::Legacy);
+ scatterNodesFromRecvBufferYGPU27AfterFtoC(para.get(), level, CudaStreamIndex::Legacy);
// Z
- prepareExchangeCollDataZGPU27AfterFtoC(para.get(), level, -1);
- exchangeCollDataZGPU27AfterFtoC(para.get(), comm, cudaMemoryManager.get(), level, -1);
- scatterNodesFromRecvBufferZGPU27AfterFtoC(para.get(), level, -1);
+ prepareExchangeCollDataZGPU27AfterFtoC(para.get(), level, CudaStreamIndex::Legacy);
+ exchangeCollDataZGPU27AfterFtoC(para.get(), comm, cudaMemoryManager.get(), level, CudaStreamIndex::Legacy);
+ scatterNodesFromRecvBufferZGPU27AfterFtoC(para.get(), level, CudaStreamIndex::Legacy);
} else {
// X
- prepareExchangeCollDataXGPU27AllNodes(para.get(), level, -1);
- exchangeCollDataXGPU27AllNodes(para.get(), comm, cudaMemoryManager.get(), level, -1);
- scatterNodesFromRecvBufferXGPU27AllNodes(para.get(), level, -1);
+ prepareExchangeCollDataXGPU27AllNodes(para.get(), level, CudaStreamIndex::Legacy);
+ exchangeCollDataXGPU27AllNodes(para.get(), comm, cudaMemoryManager.get(), level, CudaStreamIndex::Legacy);
+ scatterNodesFromRecvBufferXGPU27AllNodes(para.get(), level, CudaStreamIndex::Legacy);
// Y
- prepareExchangeCollDataYGPU27AllNodes(para.get(), level, -1);
- exchangeCollDataYGPU27AllNodes(para.get(), comm, cudaMemoryManager.get(), level, -1);
- scatterNodesFromRecvBufferYGPU27AllNodes(para.get(), level, -1);
+ prepareExchangeCollDataYGPU27AllNodes(para.get(), level, CudaStreamIndex::Legacy);
+ exchangeCollDataYGPU27AllNodes(para.get(), comm, cudaMemoryManager.get(), level, CudaStreamIndex::Legacy);
+ scatterNodesFromRecvBufferYGPU27AllNodes(para.get(), level, CudaStreamIndex::Legacy);
// Z
- prepareExchangeCollDataZGPU27AllNodes(para.get(), level, -1);
- exchangeCollDataZGPU27AllNodes(para.get(), comm, cudaMemoryManager.get(), level, -1);
- scatterNodesFromRecvBufferZGPU27AllNodes(para.get(), level, -1);
+ prepareExchangeCollDataZGPU27AllNodes(para.get(), level, CudaStreamIndex::Legacy);
+ exchangeCollDataZGPU27AllNodes(para.get(), comm, cudaMemoryManager.get(), level, CudaStreamIndex::Legacy);
+ scatterNodesFromRecvBufferZGPU27AllNodes(para.get(), level, CudaStreamIndex::Legacy);
}
//////////////////////////////////////////////////////////////////////////
@@ -204,7 +210,7 @@ void UpdateGrid27::exchangeMultiGPU_noStreams_withPrepare(int level, bool useRed
exchangePostCollDataADZGPU27(para.get(), comm, cudaMemoryManager.get(), level);
}
}
-void UpdateGrid27::exchangeMultiGPUAfterFtoC(int level, int streamIndex)
+void UpdateGrid27::exchangeMultiGPUAfterFtoC(int level, CudaStreamIndex streamIndex)
{
//////////////////////////////////////////////////////////////////////////
// 3D domain decomposition
@@ -227,9 +233,10 @@ void UpdateGrid27::exchangeMultiGPUAfterFtoC(int level, int streamIndex)
}
}
-void UpdateGrid27::postCollisionBC(int level)
+void UpdateGrid27::postCollisionBC(int level, uint t)
{
//////////////////////////////////////////////////////////////////////////
+ // G E O M E T R Y
// V E L O C I T Y (I N F L O W)
this->bcKernelManager->runVelocityBCKernelPost(level);
@@ -257,6 +264,10 @@ void UpdateGrid27::postCollisionBC(int level)
// P R E S S U R E
this->bcKernelManager->runPressureBCKernelPost(level);
+ //////////////////////////////////////////////////////////////////////////
+ // P R E C U R S O R
+ this->bcKernelManager->runPrecursorBCKernelPost(level, t, cudaMemoryManager.get());
+
//////////////////////////////////////////////////////////////////////////
// A D V E C T I O N D I F F U S I O N
if (para->getDiffOn())
@@ -317,13 +328,12 @@ void UpdateGrid27::preCollisionBC(int level, unsigned int t)
//////////////////////////////////////////////////////////////////////////////////
}
-void UpdateGrid27::fineToCoarse(int level, InterpolationCellFC* icellFC, OffFC &offFC,
- int streamIndex)
+void UpdateGrid27::fineToCoarse(int level, InterpolationCellFC* icellFC, OffFC &offFC, CudaStreamIndex streamIndex)
{
gridScalingKernelManager->runFineToCoarseKernelLB(level, icellFC, offFC, streamIndex);
if (para->getDiffOn()) {
- if (streamIndex != -1) {
+ if (para->getStreamManager()->streamIsRegistered(streamIndex)) {
printf("fineToCoarse Advection Diffusion not implemented"); // TODO
return;
}
@@ -331,14 +341,13 @@ void UpdateGrid27::fineToCoarse(int level, InterpolationCellFC* icellFC, OffFC &
}
}
-void UpdateGrid27::coarseToFine(int level, InterpolationCellCF* icellCF, OffCF &offCF,
- int streamIndex)
+void UpdateGrid27::coarseToFine(int level, InterpolationCellCF* icellCF, OffCF &offCF, CudaStreamIndex streamIndex)
{
this->gridScalingKernelManager->runCoarseToFineKernelLB(level, icellCF, offCF, streamIndex);
if (para->getDiffOn())
{
- if (streamIndex != -1){
+ if(para->getStreamManager()->streamIsRegistered(streamIndex)){
printf("CoarseToFineWithStream Advection Diffusion not implemented"); // TODO
return;
}
diff --git a/src/gpu/VirtualFluids_GPU/Calculation/UpdateGrid27.h b/src/gpu/VirtualFluids_GPU/Calculation/UpdateGrid27.h
index 8110923bf066412e2bb09ffa1f10efe3ddc983c7..8ce2cf5bfd72f9f53cdb35bc92502ee9ca0d3ad8 100644
--- a/src/gpu/VirtualFluids_GPU/Calculation/UpdateGrid27.h
+++ b/src/gpu/VirtualFluids_GPU/Calculation/UpdateGrid27.h
@@ -4,6 +4,7 @@
#include "LBM/LB.h"
#include "GPU/GPU_Interface.h"
#include "Parameter/Parameter.h"
+#include "Parameter/CudaStreamManager.h"
#include "GPU/CudaMemoryManager.h"
#include "Communication/Communicator.h"
#include "Calculation/PorousMedia.h"
@@ -15,7 +16,6 @@ class Kernel;
class BoundaryConditionFactory;
class GridScalingFactory;
class TurbulenceModelFactory;
-
class UpdateGrid27;
using CollisionStrategy = std::function<void (UpdateGrid27* updateGrid, Parameter* para, int level, unsigned int t)>;
using RefinementStrategy = std::function<void (UpdateGrid27* updateGrid, Parameter* para, int level)>;
@@ -31,21 +31,21 @@ public:
private:
void collisionAllNodes(int level, unsigned int t);
- void collisionUsingIndices(int level, unsigned int t, uint *fluidNodeIndices = nullptr, uint numberOfFluidNodes = 0, int stream = -1);
+ void collisionUsingIndices(int level, unsigned int t, uint *taggedFluidNodeIndices = nullptr, uint numberOfTaggedFluidNodes = 0, CollisionTemplate collisionTemplate = CollisionTemplate::Default, CudaStreamIndex streamIndex=CudaStreamIndex::Legacy);
void collisionAdvectionDiffusion(int level);
- void postCollisionBC(int level);
+ void postCollisionBC(int level, unsigned int t);
void preCollisionBC(int level, unsigned int t);
void collisionPorousMedia(int level);
- void fineToCoarse(int level, InterpolationCellFC* icellFC, OffFC &offFC, int streamIndex);
- void coarseToFine(int level, InterpolationCellCF* icellCF, OffCF &offCF, int streamIndex);
+ void fineToCoarse(int level, InterpolationCellFC* icellFC, OffFC &offFC, CudaStreamIndex streamIndex);
+ void coarseToFine(int level, InterpolationCellCF* icellCF, OffCF &offCF, CudaStreamIndex streamIndex);
- void prepareExchangeMultiGPU(int level, int streamIndex);
- void prepareExchangeMultiGPUAfterFtoC(int level, int streamIndex);
+ void prepareExchangeMultiGPU(int level, CudaStreamIndex streamIndex);
+ void prepareExchangeMultiGPUAfterFtoC(int level, CudaStreamIndex streamIndex);
- void exchangeMultiGPU(int level, int streamIndex);
- void exchangeMultiGPUAfterFtoC(int level, int streamIndex);
+ void exchangeMultiGPU(int level, CudaStreamIndex streamIndex);
+ void exchangeMultiGPUAfterFtoC(int level, CudaStreamIndex streamIndex);
void exchangeMultiGPU_noStreams_withPrepare(int level, bool useReducedComm);
void swapBetweenEvenAndOddTimestep(int level);
@@ -60,6 +60,7 @@ private:
friend class CollisionAndExchange_noStreams_indexKernel;
friend class CollisionAndExchange_noStreams_oldKernel;
friend class CollisionAndExchange_streams;
+ friend class CollisionAndExchange_noStreams_withReadWriteFlags;
RefinementStrategy refinement;
friend class RefinementAndExchange_streams_exchangeInterface;
diff --git a/src/gpu/VirtualFluids_GPU/Communication/ExchangeData27.cpp b/src/gpu/VirtualFluids_GPU/Communication/ExchangeData27.cpp
index 36c250401e0775b3abcc7d25c0f89fde0556631e..00a7b45668e2050467f3d1122455dc74d0ad4f1c 100644
--- a/src/gpu/VirtualFluids_GPU/Communication/ExchangeData27.cpp
+++ b/src/gpu/VirtualFluids_GPU/Communication/ExchangeData27.cpp
@@ -11,12 +11,12 @@ using namespace vf::lbm::dir;
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// 3D domain decomposition: functions used by all directions
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-void collectNodesInSendBufferGPU(Parameter *para, int level, int streamIndex,
+void collectNodesInSendBufferGPU(Parameter *para, int level, CudaStreamIndex streamIndex,
std::vector<ProcessNeighbor27> *sendProcessNeighbor,
unsigned int numberOfSendProcessNeighbors)
{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
-
+ cudaStream_t stream = para->getStreamManager()->getStream(streamIndex);
+
for (unsigned int i = 0; i < numberOfSendProcessNeighbors; i++) {
GetSendFsPostDev27(para->getParD(level)->distributions.f[0],
(*sendProcessNeighbor)[i].f[0],
@@ -32,11 +32,11 @@ void collectNodesInSendBufferGPU(Parameter *para, int level, int streamIndex,
}
}
-void scatterNodesFromRecvBufferGPU(Parameter *para, int level, int streamIndex,
+void scatterNodesFromRecvBufferGPU(Parameter *para, int level, CudaStreamIndex streamIndex,
std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
unsigned int numberOfRecvProcessNeighbors)
{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
+ cudaStream_t stream = para->getStreamManager()->getStream(streamIndex);
for (unsigned int i = 0; i < numberOfRecvProcessNeighbors; i++) {
SetRecvFsPostDev27(para->getParD(level)->distributions.f[0],
(*recvProcessNeighborDev)[i].f[0],
@@ -105,22 +105,22 @@ void copyEdgeNodes(std::vector<LBMSimulationParameter::EdgeNodePositions> &edgeN
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// X
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-void prepareExchangeCollDataXGPU27AllNodes(Parameter *para, int level, int streamIndex)
+void prepareExchangeCollDataXGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex)
{
collectNodesInSendBufferGPU(para, level, streamIndex, ¶->getParD(level)->sendProcessNeighborX,
(unsigned int)(para->getNumberOfProcessNeighborsX(level, "send")));
}
-void prepareExchangeCollDataXGPU27AfterFtoC(Parameter *para, int level, int streamIndex)
+void prepareExchangeCollDataXGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex)
{
collectNodesInSendBufferGPU(para, level, streamIndex, ¶->getParD(level)->sendProcessNeighborsAfterFtoCX,
(unsigned int)(para->getNumberOfProcessNeighborsX(level, "send")));
}
void exchangeCollDataXGPU27AllNodes(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex)
+ int level, CudaStreamIndex streamIndex)
{
- exchangeCollDataXGPU27(para, comm, cudaMemoryManager, level, streamIndex,
+ exchangeCollDataXGPU27(para, comm, cudaMemoryManager, level, streamIndex,
¶->getParD(level)->sendProcessNeighborX,
¶->getParD(level)->recvProcessNeighborX,
¶->getParH(level)->sendProcessNeighborX,
@@ -128,40 +128,40 @@ void exchangeCollDataXGPU27AllNodes(Parameter *para, vf::gpu::Communicator &comm
}
void exchangeCollDataXGPU27AfterFtoC(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex)
+ int level, CudaStreamIndex streamIndex)
{
- exchangeCollDataXGPU27(para, comm, cudaMemoryManager, level, streamIndex,
+ exchangeCollDataXGPU27(para, comm, cudaMemoryManager, level, streamIndex,
¶->getParD(level)->sendProcessNeighborsAfterFtoCX,
¶->getParD(level)->recvProcessNeighborsAfterFtoCX,
¶->getParH(level)->sendProcessNeighborsAfterFtoCX,
¶->getParH(level)->recvProcessNeighborsAfterFtoCX);
}
-void scatterNodesFromRecvBufferXGPU27AllNodes(Parameter *para, int level, int streamIndex)
+void scatterNodesFromRecvBufferXGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex)
{
- scatterNodesFromRecvBufferGPU(para, level, streamIndex, ¶->getParD(level)->recvProcessNeighborX,
+ scatterNodesFromRecvBufferGPU(para, level, streamIndex,¶->getParD(level)->recvProcessNeighborX,
(unsigned int)(para->getNumberOfProcessNeighborsX(level, "send")));
}
-void scatterNodesFromRecvBufferXGPU27AfterFtoC(Parameter *para, int level, int streamIndex)
+void scatterNodesFromRecvBufferXGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex)
{
- scatterNodesFromRecvBufferGPU(para, level, streamIndex, ¶->getParD(level)->recvProcessNeighborsAfterFtoCX,
+ scatterNodesFromRecvBufferGPU(para, level, streamIndex,¶->getParD(level)->recvProcessNeighborsAfterFtoCX,
(unsigned int)(para->getNumberOfProcessNeighborsX(level, "send")));
}
-void exchangeCollDataXGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager, int level,
- int streamIndex,
+void exchangeCollDataXGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
+ int level, CudaStreamIndex streamIndex,
std::vector<ProcessNeighbor27> *sendProcessNeighborDev,
std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
std::vector<ProcessNeighbor27> *sendProcessNeighborHost,
std::vector<ProcessNeighbor27> *recvProcessNeighborHost)
{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
+ cudaStream_t stream = para->getStreamManager()->getStream(streamIndex);
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
//! \details steps:
//! 1. copy data from device to host
for (unsigned int i = 0; i < (unsigned int)(para->getNumberOfProcessNeighborsX(level, "send")); i++)
- cudaMemoryManager->cudaCopyProcessNeighborXFsDH(level, i, (*sendProcessNeighborDev)[i].memsizeFs, streamIndex);
+ cudaMemoryManager->cudaCopyProcessNeighborXFsDH(level, i, (*sendProcessNeighborDev)[i].memsizeFs);
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
//! 2. start non-blocking receive (MPI)
@@ -181,7 +181,7 @@ void exchangeCollDataXGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMe
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
//! 7. copy received data from host to device
for (unsigned int i = 0; i < (unsigned int)(para->getNumberOfProcessNeighborsX(level, "send")); i++)
- cudaMemoryManager->cudaCopyProcessNeighborXFsHD(level, i, (*recvProcessNeighborDev)[i].memsizeFs, streamIndex);
+ cudaMemoryManager->cudaCopyProcessNeighborXFsHD(level, i, (*recvProcessNeighborDev)[i].memsizeFs);
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
}
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
@@ -189,22 +189,22 @@ void exchangeCollDataXGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMe
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// Y
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-void prepareExchangeCollDataYGPU27AllNodes(Parameter *para, int level, int streamIndex)
+void prepareExchangeCollDataYGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex)
{
collectNodesInSendBufferGPU(para, level, streamIndex, ¶->getParD(level)->sendProcessNeighborY,
(unsigned int)(para->getNumberOfProcessNeighborsY(level, "send")));
}
-void prepareExchangeCollDataYGPU27AfterFtoC(Parameter *para, int level, int streamIndex)
+void prepareExchangeCollDataYGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex)
{
collectNodesInSendBufferGPU(para, level, streamIndex, ¶->getParD(level)->sendProcessNeighborsAfterFtoCY,
(unsigned int)(para->getNumberOfProcessNeighborsY(level, "send")));
}
void exchangeCollDataYGPU27AllNodes(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex)
+ int level, CudaStreamIndex streamIndex)
{
- exchangeCollDataYGPU27(para, comm, cudaMemoryManager, level, streamIndex,
+ exchangeCollDataYGPU27(para, comm, cudaMemoryManager, level, streamIndex,
¶->getParD(level)->sendProcessNeighborY,
¶->getParD(level)->recvProcessNeighborY,
¶->getParH(level)->sendProcessNeighborY,
@@ -212,38 +212,39 @@ void exchangeCollDataYGPU27AllNodes(Parameter *para, vf::gpu::Communicator &comm
}
void exchangeCollDataYGPU27AfterFtoC(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex)
+ int level, CudaStreamIndex streamIndex)
{
- exchangeCollDataYGPU27(para, comm, cudaMemoryManager, level, streamIndex,
+ exchangeCollDataYGPU27(para, comm, cudaMemoryManager, level, streamIndex,
¶->getParD(level)->sendProcessNeighborsAfterFtoCY,
¶->getParD(level)->recvProcessNeighborsAfterFtoCY,
¶->getParH(level)->sendProcessNeighborsAfterFtoCY,
¶->getParH(level)->recvProcessNeighborsAfterFtoCY);
}
-void scatterNodesFromRecvBufferYGPU27AllNodes(Parameter *para, int level, int streamIndex)
+void scatterNodesFromRecvBufferYGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex)
{
scatterNodesFromRecvBufferGPU(para, level, streamIndex, ¶->getParD(level)->recvProcessNeighborY,
(unsigned int)(para->getNumberOfProcessNeighborsY(level, "send")));
}
-void scatterNodesFromRecvBufferYGPU27AfterFtoC(Parameter *para, int level, int streamIndex)
+void scatterNodesFromRecvBufferYGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex)
{
scatterNodesFromRecvBufferGPU(para, level, streamIndex, ¶->getParD(level)->recvProcessNeighborsAfterFtoCY,
(unsigned int)(para->getNumberOfProcessNeighborsY(level, "send")));
}
void exchangeCollDataYGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager, int level,
- int streamIndex, std::vector<ProcessNeighbor27> *sendProcessNeighborDev,
+ CudaStreamIndex streamIndex,
+ std::vector<ProcessNeighbor27> *sendProcessNeighborDev,
std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
std::vector<ProcessNeighbor27> *sendProcessNeighborHost,
std::vector<ProcessNeighbor27> *recvProcessNeighborHost)
{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
+ cudaStream_t stream = para->getStreamManager()->getStream(CudaStreamIndex::SubDomainBorder);
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// copy Device to Host
for (unsigned int i = 0; i < (unsigned int)(para->getNumberOfProcessNeighborsY(level, "send")); i++)
- cudaMemoryManager->cudaCopyProcessNeighborYFsDH(level, i, (*sendProcessNeighborDev)[i].memsizeFs, streamIndex);
+ cudaMemoryManager->cudaCopyProcessNeighborYFsDH(level, i, (*sendProcessNeighborDev)[i].memsizeFs);
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
startNonBlockingMpiReceive((unsigned int)(*sendProcessNeighborHost).size(), comm, recvProcessNeighborHost);
@@ -276,7 +277,7 @@ void exchangeCollDataYGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMe
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// copy Host to Device
for (unsigned int i = 0; i < (unsigned int)(para->getNumberOfProcessNeighborsY(level, "send")); i++) {
- cudaMemoryManager->cudaCopyProcessNeighborYFsHD(level, i, (*recvProcessNeighborDev)[i].memsizeFs, streamIndex);
+ cudaMemoryManager->cudaCopyProcessNeighborYFsHD(level, i, (*recvProcessNeighborDev)[i].memsizeFs);
}
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
}
@@ -285,61 +286,62 @@ void exchangeCollDataYGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMe
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// Z
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-void prepareExchangeCollDataZGPU27AllNodes(Parameter *para, int level, int streamIndex)
+void prepareExchangeCollDataZGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex)
{
collectNodesInSendBufferGPU(para, level, streamIndex, ¶->getParD(level)->sendProcessNeighborZ,
(unsigned int)(para->getNumberOfProcessNeighborsZ(level, "send")));
}
-void prepareExchangeCollDataZGPU27AfterFtoC(Parameter *para, int level, int streamIndex)
+void prepareExchangeCollDataZGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex)
{
collectNodesInSendBufferGPU(para, level, streamIndex, ¶->getParD(level)->sendProcessNeighborsAfterFtoCZ,
(unsigned int)(para->getNumberOfProcessNeighborsZ(level, "send")));
}
void exchangeCollDataZGPU27AllNodes(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex)
+ int level, CudaStreamIndex streamIndex)
{
- exchangeCollDataZGPU27(para, comm, cudaMemoryManager, level, streamIndex,
+ exchangeCollDataZGPU27(para, comm, cudaMemoryManager, level, streamIndex,
¶->getParD(level)->sendProcessNeighborZ,
¶->getParD(level)->recvProcessNeighborZ,
¶->getParH(level)->sendProcessNeighborZ,
¶->getParH(level)->recvProcessNeighborZ);
}
void exchangeCollDataZGPU27AfterFtoC(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex)
+ int level, CudaStreamIndex streamIndex)
{
- exchangeCollDataZGPU27(para, comm, cudaMemoryManager, level, streamIndex,
+ exchangeCollDataZGPU27(para, comm, cudaMemoryManager, level, streamIndex,
¶->getParD(level)->sendProcessNeighborsAfterFtoCZ,
¶->getParD(level)->recvProcessNeighborsAfterFtoCZ,
¶->getParH(level)->sendProcessNeighborsAfterFtoCZ,
¶->getParH(level)->recvProcessNeighborsAfterFtoCZ);
}
-void scatterNodesFromRecvBufferZGPU27AllNodes(Parameter *para, int level, int streamIndex)
+void scatterNodesFromRecvBufferZGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex)
{
scatterNodesFromRecvBufferGPU(para, level, streamIndex, ¶->getParD(level)->recvProcessNeighborZ,
(unsigned int)(para->getNumberOfProcessNeighborsZ(level, "send")));
}
-void scatterNodesFromRecvBufferZGPU27AfterFtoC(Parameter *para, int level, int streamIndex)
+void scatterNodesFromRecvBufferZGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex)
{
scatterNodesFromRecvBufferGPU(para, level, streamIndex, ¶->getParD(level)->recvProcessNeighborsAfterFtoCZ,
(unsigned int)(para->getNumberOfProcessNeighborsZ(level, "send")));
}
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-void exchangeCollDataZGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager, int level,
- int streamIndex, std::vector<ProcessNeighbor27> *sendProcessNeighborDev,
+void exchangeCollDataZGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager, int level,
+ CudaStreamIndex streamIndex,
+ std::vector<ProcessNeighbor27> *sendProcessNeighborDev,
std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
std::vector<ProcessNeighbor27> *sendProcessNeighborHost,
std::vector<ProcessNeighbor27> *recvProcessNeighborHost)
{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
+ cudaStream_t stream = para->getStreamManager()->getStream(streamIndex);
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// copy Device to Host
for (unsigned int i = 0; i < (unsigned int)(para->getNumberOfProcessNeighborsZ(level, "send")); i++)
- cudaMemoryManager->cudaCopyProcessNeighborZFsDH(level, i, (*sendProcessNeighborDev)[i].memsizeFs, streamIndex);
+ cudaMemoryManager->cudaCopyProcessNeighborZFsDH(level, i, (*sendProcessNeighborDev)[i].memsizeFs);
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
startNonBlockingMpiReceive((unsigned int)(*sendProcessNeighborHost).size(), comm, recvProcessNeighborHost);
/////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
@@ -386,7 +388,7 @@ void exchangeCollDataZGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMe
// copy Host to Device
for (unsigned int i = 0; i < (unsigned int)(para->getNumberOfProcessNeighborsZ(level, "send")); i++)
{
- cudaMemoryManager->cudaCopyProcessNeighborZFsHD(level, i, (*recvProcessNeighborDev)[i].memsizeFs, streamIndex);
+ cudaMemoryManager->cudaCopyProcessNeighborZFsHD(level, i, (*recvProcessNeighborDev)[i].memsizeFs);
}
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
}
diff --git a/src/gpu/VirtualFluids_GPU/Communication/ExchangeData27.h b/src/gpu/VirtualFluids_GPU/Communication/ExchangeData27.h
index ec930ebbc06554e948204b74e79e0e25b85f57b5..8302ffdc47bfa012c47df00f90c2491039f4eaee 100644
--- a/src/gpu/VirtualFluids_GPU/Communication/ExchangeData27.h
+++ b/src/gpu/VirtualFluids_GPU/Communication/ExchangeData27.h
@@ -6,6 +6,7 @@
#include "GPU/GPU_Interface.h"
#include "LBM/LB.h"
#include "Parameter/Parameter.h"
+#include "Parameter/CudaStreamManager.h"
//! \file ExchangeData27.h
//! \ingroup GPU
@@ -14,9 +15,9 @@
//////////////////////////////////////////////////////////////////////////
// 1D domain decomposition
-void exchangePreCollDataGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
+void exchangePreCollDataGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
int level);
-void exchangePostCollDataGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
+void exchangePostCollDataGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
int level);
//////////////////////////////////////////////////////////////////////////
// 3D domain decomposition
@@ -24,13 +25,13 @@ void exchangePostCollDataGPU27(Parameter *para, vf::gpu::Communicator &comm, Cud
// functions used for all directions
//! \brief Collect the send nodes in a buffer on the gpu
-void collectNodesInSendBufferGPU(Parameter *para, int level, int streamIndex,
- std::vector<ProcessNeighbor27> *sendProcessNeighbor,
- unsigned int numberOfSendProcessNeighbors);
+void collectNodesInSendBufferGPU(Parameter *para, int level, CudaStreamIndex streamIndex,
+ std::vector<ProcessNeighbor27> *sendProcessNeighbor,
+ unsigned int numberOfSendProcessNeighbors);
//! \brief Distribute the receive nodes from the buffer on the gpu
-void scatterNodesFromRecvBufferGPU(Parameter *para, int level, int streamIndex,
- std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
- unsigned int numberOfRecvProcessNeighbors);
+void scatterNodesFromRecvBufferGPU(Parameter *para, int level, CudaStreamIndex streamIndex,
+ std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
+ unsigned int numberOfRecvProcessNeighbors);
//! \brief Copy nodes which are part of the communication in multiple directions
//! \details The nodes are copied from the receive buffer in one direction to the send buffer in another direction. The
//! copy operation is conducted on the cpu.
@@ -49,21 +50,20 @@ void copyEdgeNodes(std::vector<LBMSimulationParameter::EdgeNodePositions> &edgeN
//! \brief Collect the send nodes for communication in the x direction in a buffer on the gpu
//! \details Needed to exchange all nodes, used in the communication after collision step
-void prepareExchangeCollDataXGPU27AllNodes(Parameter *para, int level, int streamIndex);
+void prepareExchangeCollDataXGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex);
//! \brief Collect the send nodes for communication in the x direction in a buffer on the gpu
//! \details Only exchange nodes which are part of the interpolation process on refined grids. This function is used in
//! the exchange which takes place after the interpolation fine to coarse and before the interpolation coarse to fine.
//! See [master thesis of Anna Wellmann]
-void prepareExchangeCollDataXGPU27AfterFtoC(Parameter *para, int level, int streamIndex);
+void prepareExchangeCollDataXGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex);
//! \brief Exchange routine in x direction for simulations on multiple gpus
//! \details Send and receive the nodes from the communication buffers on the gpus.
//! \param Communicator is needed for the communication between the processes with mpi
//! \param CudaMemoryManager is needed for moving the data between host and device
-//! \param streamIndex is the index of a CUDA Stream, which is needed for communication hiding
//! \param sendProcessNeighborDev, recvProcessNeighborDev, sendProcessNeighborHost, recvProcessNeighborHost are pointers
//! to the send and receive arrays, both on the device and the host
void exchangeCollDataXGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex,
+ int level, CudaStreamIndex streamIndex,
std::vector<ProcessNeighbor27> *sendProcessNeighborDev,
std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
std::vector<ProcessNeighbor27> *sendProcessNeighborHost,
@@ -71,59 +71,59 @@ void exchangeCollDataXGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMe
//! \brief Calls exchangeCollDataXGPU27() for exchanging all nodes
//! \details Used in the communication after collision step
void exchangeCollDataXGPU27AllNodes(Parameter *para, vf::gpu::Communicator &comm,
- CudaMemoryManager *cudaMemoryManager, int level, int streamIndex);
+ CudaMemoryManager *cudaMemoryManager, int level, CudaStreamIndex streamIndex);
//! \brief Calls exchangeCollDataXGPU27() for exchanging the nodes, which are part of the communication between the two
//! interpolation processes on refined grids
//! \details Only exchange nodes which are part of the interpolation process on
//! refined grids. This function is used in the exchange which takes place after the interpolation fine to coarse and
//! before the interpolation coarse to fine. See [master thesis of Anna Wellmann]
void exchangeCollDataXGPU27AfterFtoC(Parameter *para, vf::gpu::Communicator &comm,
- CudaMemoryManager *cudaMemoryManager, int level, int streamIndex);
+ CudaMemoryManager *cudaMemoryManager, int level, CudaStreamIndex streamIndex);
//! \brief Distribute the receive nodes (x direction) from the buffer on the gpu
//! \details Needed to exchange all nodes, used in the communication after collision step
-void scatterNodesFromRecvBufferXGPU27AllNodes(Parameter *para, int level, int streamIndex);
+void scatterNodesFromRecvBufferXGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex);
//! \brief Distribute the receive nodes (x direction) from the buffer on the gpu
//! \details Only exchange nodes which are part of the interpolation process on refined grids. This function is used in
//! the exchange which takes place after the interpolation fine to coarse and before the interpolation coarse to fine.
//! See [master thesis of Anna Wellmann]
-void scatterNodesFromRecvBufferXGPU27AfterFtoC(Parameter *para, int level, int streamIndex);
+void scatterNodesFromRecvBufferXGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex);
//////////////////////////////////////////////////////////////////////////
// y
-void prepareExchangeCollDataYGPU27AllNodes(Parameter *para, int level, int streamIndex);
-void prepareExchangeCollDataYGPU27AfterFtoC(Parameter *para, int level, int streamIndex);
+void prepareExchangeCollDataYGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex);
+void prepareExchangeCollDataYGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex);
void exchangeCollDataYGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex,
+ int level,CudaStreamIndex streamIndex,
std::vector<ProcessNeighbor27> *sendProcessNeighborDev,
std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
std::vector<ProcessNeighbor27> *sendProcessNeighborHost,
std::vector<ProcessNeighbor27> *recvProcessNeighborHos);
void exchangeCollDataYGPU27AllNodes(Parameter *para, vf::gpu::Communicator &comm,
- CudaMemoryManager *cudaMemoryManager, int level, int streamIndex);
+ CudaMemoryManager *cudaMemoryManager, int level, CudaStreamIndex streamIndex);
void exchangeCollDataYGPU27AfterFtoC(Parameter *para, vf::gpu::Communicator &comm,
- CudaMemoryManager *cudaMemoryManager, int level, int streamIndex);
-void scatterNodesFromRecvBufferYGPU27AllNodes(Parameter *para, int level, int streamIndex);
-void scatterNodesFromRecvBufferYGPU27AfterFtoC(Parameter *para, int level, int streamIndex);
+ CudaMemoryManager *cudaMemoryManager, int level, CudaStreamIndex streamIndex);
+void scatterNodesFromRecvBufferYGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex);
+void scatterNodesFromRecvBufferYGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex);
// z
-void prepareExchangeCollDataZGPU27AllNodes(Parameter *para, int level, int streamIndex);
-void prepareExchangeCollDataZGPU27AfterFtoC(Parameter *para, int level, int streamIndex);
+void prepareExchangeCollDataZGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex);
+void prepareExchangeCollDataZGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex);
void exchangeCollDataZGPU27(Parameter *para, vf::gpu::Communicator &comm, CudaMemoryManager *cudaMemoryManager,
- int level, int streamIndex,
+ int level, CudaStreamIndex streamIndex,
std::vector<ProcessNeighbor27> *sendProcessNeighborDev,
std::vector<ProcessNeighbor27> *recvProcessNeighborDev,
std::vector<ProcessNeighbor27> *sendProcessNeighborHost,
std::vector<ProcessNeighbor27> *recvProcessNeighborHost);
void exchangeCollDataZGPU27AllNodes(Parameter *para, vf::gpu::Communicator &comm,
- CudaMemoryManager *cudaMemoryManager, int level, int streamIndex);
+ CudaMemoryManager *cudaMemoryManager, int level, CudaStreamIndex streamIndex);
void exchangeCollDataZGPU27AfterFtoC(Parameter *para, vf::gpu::Communicator &comm,
- CudaMemoryManager *cudaMemoryManager, int level, int streamIndex);
+ CudaMemoryManager *cudaMemoryManager, int level, CudaStreamIndex streamIndex);
-void scatterNodesFromRecvBufferZGPU27AllNodes(Parameter *para, int level, int streamIndex);
-void scatterNodesFromRecvBufferZGPU27AfterFtoC(Parameter *para, int level, int streamIndex);
+void scatterNodesFromRecvBufferZGPU27AllNodes(Parameter *para, int level, CudaStreamIndex streamIndex);
+void scatterNodesFromRecvBufferZGPU27AfterFtoC(Parameter *para, int level, CudaStreamIndex streamIndex);
//////////////////////////////////////////////////////////////////////////
// 3D domain decomposition convection diffusion
diff --git a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.cpp b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.cpp
index e197fb5c28611e77406b30ab39aa6af2f54b9ef5..c996525ee78e6ba87fe0972df084aedccfefd39a 100644
--- a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.cpp
+++ b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.cpp
@@ -29,15 +29,10 @@ void GridProvider::setNumberOfNodes(const int numberOfNodes, const int level) co
para->getParD(level)->mem_size_int_SP = sizeof(uint) * para->getParD(level)->numberOfNodes;
}
-void GridProvider::setNumberOfFluidNodes(const int numberOfNodes, const int level) const
+void GridProvider::setNumberOfTaggedFluidNodes(const int numberOfNodes, CollisionTemplate tag, const int level) const
{
- para->getParH(level)->numberOfFluidNodes = numberOfNodes;
- para->getParD(level)->numberOfFluidNodes = numberOfNodes;
-}
-
-void GridProvider::setNumberOfFluidNodesBorder(const int numberOfNodes, const int level) const {
- para->getParH(level)->numberOfFluidNodesBorder = numberOfNodes;
- para->getParD(level)->numberOfFluidNodesBorder = numberOfNodes;
+ para->getParH(level)->numberOfTaggedFluidNodes[tag] = numberOfNodes;
+ para->getParD(level)->numberOfTaggedFluidNodes[tag] = numberOfNodes;
}
void GridProvider::setInitalNodeValues(const int numberOfNodes, const int level) const
diff --git a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.h b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.h
index 5fc5826735643ec748da169160e782004d7e5fb7..42ba7a464a6e76fd747cc6d11a01e3957c865765 100644
--- a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.h
+++ b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridProvider.h
@@ -5,7 +5,7 @@
#include <vector>
#include <memory>
-
+#include "LBM/LB.h"
#include "PointerDefinitions.h"
#include "VirtualFluids_GPU_export.h"
#include "gpu/GridGenerator/io/SimulationFileWriter/SimulationFileWriter.h"
@@ -28,8 +28,10 @@ public:
virtual void allocArrays_BoundaryValues() = 0;
virtual void allocArrays_BoundaryQs() = 0;
virtual void allocArrays_OffsetScale() = 0;
- virtual void allocArrays_fluidNodeIndices() = 0;
- virtual void allocArrays_fluidNodeIndicesBorder() = 0;
+ virtual void allocArrays_taggedFluidNodes() = 0;
+
+ virtual void tagFluidNodeIndices(std::vector<uint> taggedFluidNodeIndices, CollisionTemplate tag, uint level) = 0;
+ virtual void sortFluidNodeTags() = 0;
virtual void setDimensions() = 0;
virtual void setBoundingBox() = 0;
@@ -45,8 +47,7 @@ public:
protected:
void setNumberOfNodes(const int numberOfNodes, const int level) const;
- void setNumberOfFluidNodes(const int numberOfNodes, const int level) const;
- void setNumberOfFluidNodesBorder(const int numberOfNodes, const int level) const;
+ void setNumberOfTaggedFluidNodes(const int numberOfNodes, CollisionTemplate tag, const int level) const;
virtual void setInitalNodeValues(const int numberOfNodes, const int level) const;
void setPressSizePerLevel(int level, int sizePerLevel) const;
diff --git a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.cpp b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.cpp
index fa432a1d9c3922b88e93588548db74083275ef1e..793400869dd29fe12be357ffcf87c0238ea70eb9 100644
--- a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.cpp
+++ b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.cpp
@@ -218,16 +218,20 @@ void GridReader::allocArrays_OffsetScale()
std::cout << "-----Ende OffsetScale------" << std::endl;
}
-void GridReader::allocArrays_fluidNodeIndices() {
+void GridReader::allocArrays_taggedFluidNodes() {
std::cout << "GridReader::allocArrays_fluidNodeIndices not implemented" << std::endl;
// TODO
}
-void GridReader::allocArrays_fluidNodeIndicesBorder() {
- std::cout << "GridReader::allocArrays_fluidNodeIndicesBorder not implemented" << std::endl;
+void GridReader::tagFluidNodeIndices(std::vector<uint> taggedFluidNodeIndices, CollisionTemplate tag, uint level){
+ std::cout << "GridReader::tagFluidNodeIndices not implemented" << std::endl;
// TODO
}
+void GridReader::sortFluidNodeTags(){
+ std::cout << "GridReader::sortFluidNodeTags not implemented" << std::endl;
+ // TODO
+}
void GridReader::setPressureValues(int channelSide) const
{
diff --git a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.h b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.h
index 18efb6a7885191312ea4e2fbb22eb45162ab1de1..2c17c28022d88a65e922e23d1c89f5166f5a1716 100644
--- a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.h
+++ b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderFiles/GridReader.h
@@ -39,8 +39,10 @@ public:
void allocArrays_CoordNeighborGeo() override;
void allocArrays_BoundaryValues() override;
void allocArrays_OffsetScale() override;
- void allocArrays_fluidNodeIndices() override;
- void allocArrays_fluidNodeIndicesBorder() override;
+ void allocArrays_taggedFluidNodes() override;
+
+ void tagFluidNodeIndices(std::vector<uint> taggedFluidNodeIndices, CollisionTemplate tag, uint level) override;
+ void sortFluidNodeTags() override;
void initalValuesDomainDecompostion(int level);
diff --git a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderGenerator/GridGenerator.cpp b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderGenerator/GridGenerator.cpp
index 7f61b4357276f38d8fde71489dcf60348b402941..d89f4cd85985694b4529d338ab5e4c5a86cd3a38 100644
--- a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderGenerator/GridGenerator.cpp
+++ b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderGenerator/GridGenerator.cpp
@@ -1,5 +1,6 @@
#include "GridGenerator.h"
+#include "LBM/LB.h"
#include "Parameter/Parameter.h"
#include "GridGenerator/grid/GridBuilder/GridBuilder.h"
#include "GPU/CudaMemoryManager.h"
@@ -10,10 +11,13 @@
#include <algorithm>
#include "utilities/math/Math.h"
#include "Output/QDebugWriter.hpp"
+#include "GridGenerator/TransientBCSetter/TransientBCSetter.h"
#include "utilities/communication.h"
#include "Communication/Communicator.h"
+#include <logger/Logger.h>
+
using namespace vf::lbm::dir;
GridGenerator::GridGenerator(std::shared_ptr<GridBuilder> builder, std::shared_ptr<Parameter> para, std::shared_ptr<CudaMemoryManager> cudaMemoryManager, vf::gpu::Communicator& communicator):
@@ -100,22 +104,98 @@ void GridGenerator::allocArrays_CoordNeighborGeo()
std::cout << "-----finish Coord, Neighbor, Geo------" << std::endl;
}
-void GridGenerator::allocArrays_fluidNodeIndices() {
- for (uint level = 0; level < builder->getNumberOfGridLevels(); level++) {
- setNumberOfFluidNodes(builder->getNumberOfFluidNodes(level), level);
- cudaMemoryManager->cudaAllocFluidNodeIndices(level);
- builder->getFluidNodeIndices(para->getParH(level)->fluidNodeIndices, level);
- cudaMemoryManager->cudaCopyFluidNodeIndices(level);
- }
+void GridGenerator::allocArrays_taggedFluidNodes() {
+
+ for (uint level = 0; level < builder->getNumberOfGridLevels(); level++)
+ {
+ for ( CollisionTemplate tag: all_CollisionTemplate )
+ { //TODO: Need to add CollisionTemplate to GridBuilder to allow as argument and get rid of indivual get funtions for fluid node indices... and clean up this mess
+ switch(tag)
+ {
+ case CollisionTemplate::Default:
+ this->setNumberOfTaggedFluidNodes(builder->getNumberOfFluidNodes(level), CollisionTemplate::Default, level);
+ cudaMemoryManager->cudaAllocTaggedFluidNodeIndices(CollisionTemplate::Default, level);
+ builder->getFluidNodeIndices(para->getParH(level)->taggedFluidNodeIndices[CollisionTemplate::Default], level);
+ cudaMemoryManager->cudaCopyTaggedFluidNodeIndices(CollisionTemplate::Default, level);
+ if(para->getParH(level)->numberOfTaggedFluidNodes[tag]>0)
+ para->getParH(level)->allocatedBulkFluidNodeTags.push_back(tag);
+ break;
+ case CollisionTemplate::SubDomainBorder:
+ this->setNumberOfTaggedFluidNodes(builder->getNumberOfFluidNodesBorder(level), CollisionTemplate::SubDomainBorder, level);
+ cudaMemoryManager->cudaAllocTaggedFluidNodeIndices(CollisionTemplate::SubDomainBorder, level);
+ builder->getFluidNodeIndicesBorder(para->getParH(level)->taggedFluidNodeIndices[CollisionTemplate::SubDomainBorder], level);
+ cudaMemoryManager->cudaCopyTaggedFluidNodeIndices(CollisionTemplate::SubDomainBorder, level);
+ break;
+ case CollisionTemplate::WriteMacroVars:
+ this->setNumberOfTaggedFluidNodes(builder->getNumberOfFluidNodesMacroVars(level), CollisionTemplate::WriteMacroVars, level);
+ cudaMemoryManager->cudaAllocTaggedFluidNodeIndices(CollisionTemplate::WriteMacroVars, level);
+ builder->getFluidNodeIndicesMacroVars(para->getParH(level)->taggedFluidNodeIndices[CollisionTemplate::WriteMacroVars], level);
+ cudaMemoryManager->cudaCopyTaggedFluidNodeIndices(CollisionTemplate::WriteMacroVars, level);
+ if(para->getParH(level)->numberOfTaggedFluidNodes[tag]>0)
+ para->getParH(level)->allocatedBulkFluidNodeTags.push_back(tag);
+ break;
+ case CollisionTemplate::ApplyBodyForce:
+ this->setNumberOfTaggedFluidNodes(builder->getNumberOfFluidNodesApplyBodyForce(level), CollisionTemplate::ApplyBodyForce, level);
+ cudaMemoryManager->cudaAllocTaggedFluidNodeIndices(CollisionTemplate::ApplyBodyForce, level);
+ builder->getFluidNodeIndicesApplyBodyForce(para->getParH(level)->taggedFluidNodeIndices[CollisionTemplate::ApplyBodyForce], level);
+ cudaMemoryManager->cudaCopyTaggedFluidNodeIndices(CollisionTemplate::ApplyBodyForce, level);
+ if(para->getParH(level)->numberOfTaggedFluidNodes[tag]>0)
+ para->getParH(level)->allocatedBulkFluidNodeTags.push_back(tag);
+ break;
+ case CollisionTemplate::AllFeatures:
+ this->setNumberOfTaggedFluidNodes(builder->getNumberOfFluidNodesAllFeatures(level), CollisionTemplate::AllFeatures, level);
+ cudaMemoryManager->cudaAllocTaggedFluidNodeIndices(CollisionTemplate::AllFeatures, level);
+ builder->getFluidNodeIndicesAllFeatures(para->getParH(level)->taggedFluidNodeIndices[CollisionTemplate::AllFeatures], level);
+ cudaMemoryManager->cudaCopyTaggedFluidNodeIndices(CollisionTemplate::AllFeatures, level);
+ if(para->getParH(level)->numberOfTaggedFluidNodes[tag]>0)
+ para->getParH(level)->allocatedBulkFluidNodeTags.push_back(tag);
+ break;
+ default:
+ break;
+ }
+ }
+ VF_LOG_INFO("Number of tagged nodes on level {}:", level);
+ VF_LOG_INFO("Default: {}, Border: {}, WriteMacroVars: {}, ApplyBodyForce: {}, AllFeatures: {}",
+ para->getParH(level)->numberOfTaggedFluidNodes[CollisionTemplate::Default],
+ para->getParH(level)->numberOfTaggedFluidNodes[CollisionTemplate::SubDomainBorder],
+ para->getParH(level)->numberOfTaggedFluidNodes[CollisionTemplate::WriteMacroVars],
+ para->getParH(level)->numberOfTaggedFluidNodes[CollisionTemplate::ApplyBodyForce],
+ para->getParH(level)->numberOfTaggedFluidNodes[CollisionTemplate::AllFeatures] );
+ }
}
-void GridGenerator::allocArrays_fluidNodeIndicesBorder() {
- for (uint level = 0; level < builder->getNumberOfGridLevels(); level++) {
- setNumberOfFluidNodesBorder(builder->getNumberOfFluidNodesBorder(level), level);
- cudaMemoryManager->cudaAllocFluidNodeIndicesBorder(level);
- builder->getFluidNodeIndicesBorder(para->getParH(level)->fluidNodeIndicesBorder, level);
- cudaMemoryManager->cudaCopyFluidNodeIndicesBorder(level);
+void GridGenerator::tagFluidNodeIndices(std::vector<uint> taggedFluidNodeIndices, CollisionTemplate tag, uint level) {
+ switch(tag)
+ {
+ case CollisionTemplate::WriteMacroVars:
+ builder->addFluidNodeIndicesMacroVars( taggedFluidNodeIndices, level );
+ break;
+ case CollisionTemplate::ApplyBodyForce:
+ builder->addFluidNodeIndicesApplyBodyForce( taggedFluidNodeIndices, level );
+ break;
+ case CollisionTemplate::AllFeatures:
+ builder->addFluidNodeIndicesAllFeatures( taggedFluidNodeIndices, level );
+ break;
+ case CollisionTemplate::Default:
+ case CollisionTemplate::SubDomainBorder:
+ throw std::runtime_error("Cannot tag fluid nodes as Default or SubDomainBorder!");
+ default:
+ throw std::runtime_error("Tagging fluid nodes with invald tag!");
+ break;
+
+ }
+
+}
+
+void GridGenerator::sortFluidNodeTags() {
+ VF_LOG_INFO("Start sorting tagged fluid nodes...");
+ for (uint level = 0; level < builder->getNumberOfGridLevels(); level++)
+ {
+ builder->sortFluidNodeIndicesAllFeatures(level); //has to be called first!
+ builder->sortFluidNodeIndicesMacroVars(level);
+ builder->sortFluidNodeIndicesApplyBodyForce(level);
}
+ VF_LOG_INFO("done.");
}
void GridGenerator::allocArrays_BoundaryValues()
@@ -129,6 +209,7 @@ void GridGenerator::allocArrays_BoundaryValues()
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
para->getParH(level)->pressureBC.numberOfBCnodes = 0;
+ para->getParD(level)->outflowPressureCorrectionFactor = para->getOutflowPressureCorrectionFactor();
if (numberOfPressureValues > 1)
{
blocks = (numberOfPressureValues / para->getParH(level)->numberofthreads) + 1;
@@ -242,6 +323,98 @@ void GridGenerator::allocArrays_BoundaryValues()
para->getParD(level)->numberOfVeloBCnodesRead = para->getParH(level)->velocityBC.numberOfBCnodes * para->getD3Qxx();
}
+ for (uint level = 0; level < builder->getNumberOfGridLevels(); level++) {
+ const auto numberOfPrecursorValues = int(builder->getPrecursorSize(level));
+ *logging::out << logging::Logger::INFO_INTERMEDIATE << "size precursor level " << level << " : " << numberOfPrecursorValues << "\n";
+ ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ int blocks = (numberOfPrecursorValues / para->getParH(level)->numberofthreads) + 1;
+ para->getParH(level)->precursorBC.sizeQ = blocks * para->getParH(level)->numberofthreads;
+ para->getParD(level)->precursorBC.sizeQ = para->getParH(level)->precursorBC.sizeQ;
+ ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ para->getParH(level)->precursorBC.numberOfBCnodes = numberOfPrecursorValues;
+ para->getParD(level)->precursorBC.numberOfBCnodes = numberOfPrecursorValues;
+ para->getParH(level)->numberOfPrecursorBCnodesRead = numberOfPrecursorValues * para->getD3Qxx();
+ para->getParD(level)->numberOfPrecursorBCnodesRead = numberOfPrecursorValues * para->getD3Qxx();
+
+ if (numberOfPrecursorValues > 1)
+ {
+ ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ cudaMemoryManager->cudaAllocPrecursorBC(level);
+ ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ builder->getPrecursorValues(
+ para->getParH(level)->precursorBC.planeNeighbor0PP, para->getParH(level)->precursorBC.planeNeighbor0PM,
+ para->getParH(level)->precursorBC.planeNeighbor0MP, para->getParH(level)->precursorBC.planeNeighbor0MM,
+ para->getParH(level)->precursorBC.weights0PP, para->getParH(level)->precursorBC.weights0PM,
+ para->getParH(level)->precursorBC.weights0MP, para->getParH(level)->precursorBC.weights0MM,
+ para->getParH(level)->precursorBC.k, para->getParH(level)->transientBCInputFileReader, para->getParH(level)->precursorBC.numberOfPrecursorNodes,
+ para->getParH(level)->precursorBC.numberOfQuantities, para->getParH(level)->precursorBC.timeStepsBetweenReads,
+ para->getParH(level)->precursorBC.velocityX, para->getParH(level)->precursorBC.velocityY, para->getParH(level)->precursorBC.velocityZ,
+ level);
+ ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ para->getParD(level)->precursorBC.numberOfPrecursorNodes = para->getParH(level)->precursorBC.numberOfPrecursorNodes;
+ para->getParD(level)->precursorBC.numberOfQuantities = para->getParH(level)->precursorBC.numberOfQuantities;
+ para->getParD(level)->precursorBC.timeStepsBetweenReads = para->getParH(level)->precursorBC.timeStepsBetweenReads;
+ para->getParD(level)->precursorBC.velocityX = para->getParH(level)->precursorBC.velocityX;
+ para->getParD(level)->precursorBC.velocityY = para->getParH(level)->precursorBC.velocityY;
+ para->getParD(level)->precursorBC.velocityZ = para->getParH(level)->precursorBC.velocityZ;
+
+ for(auto reader : para->getParH(level)->transientBCInputFileReader)
+ {
+ if(reader->getNumberOfQuantities() != para->getParD(level)->precursorBC.numberOfQuantities) throw std::runtime_error("Number of quantities in reader and number of quantities needed for precursor don't match!");
+ }
+
+ cudaMemoryManager->cudaCopyPrecursorBC(level);
+ cudaMemoryManager->cudaAllocPrecursorData(level);
+
+ // read first timestep of precursor into next and copy to next on device
+ for(auto reader : para->getParH(level)->transientBCInputFileReader)
+ {
+ reader->getNextData(para->getParH(level)->precursorBC.next, para->getParH(level)->precursorBC.numberOfPrecursorNodes, 0);
+ }
+
+ cudaMemoryManager->cudaCopyPrecursorData(level);
+
+ //switch next with last pointers
+ real* tmp = para->getParD(level)->precursorBC.last;
+ para->getParD(level)->precursorBC.last = para->getParD(level)->precursorBC.next;
+ para->getParD(level)->precursorBC.next = tmp;
+
+ //read second timestep of precursor into next and copy next to device
+ real nextTime = para->getParD(level)->precursorBC.timeStepsBetweenReads*pow(2,-((real)level))*para->getTimeRatio();
+ for(auto reader : para->getParH(level)->transientBCInputFileReader)
+ {
+ reader->getNextData(para->getParH(level)->precursorBC.next, para->getParH(level)->precursorBC.numberOfPrecursorNodes, nextTime);
+ }
+
+ cudaMemoryManager->cudaCopyPrecursorData(level);
+
+ para->getParD(level)->precursorBC.nPrecursorReads = 1;
+
+
+ //switch next with current pointers
+ tmp = para->getParD(level)->precursorBC.current;
+ para->getParD(level)->precursorBC.current = para->getParD(level)->precursorBC.next;
+ para->getParD(level)->precursorBC.next = tmp;
+
+ //start usual cycle of loading, i.e. read velocities of timestep after current and copy asynchronously to device
+ for(auto reader : para->getParH(level)->transientBCInputFileReader)
+ {
+ reader->getNextData(para->getParH(level)->precursorBC.next, para->getParH(level)->precursorBC.numberOfPrecursorNodes, 2*nextTime);
+ }
+
+ cudaMemoryManager->cudaCopyPrecursorData(level);
+
+ para->getParD(level)->precursorBC.nPrecursorReads = 2;
+ }
+
+ ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ // advection - diffusion stuff
+ if (para->getDiffOn()==true){
+ throw std::runtime_error(" Advection Diffusion not implemented for Precursor!");
+ }
+ //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ }
+
if (builder->hasGeometryValues()) {
@@ -874,6 +1047,50 @@ void GridGenerator::allocArrays_BoundaryQs()
}
}
+ for (uint i = 0; i < builder->getNumberOfGridLevels(); i++) {
+ const auto numberOfPrecursorNodes = int(builder->getPrecursorSize(i));
+ if (numberOfPrecursorNodes > 0)
+ {
+ std::cout << "size velocity level " << i << " : " << numberOfPrecursorNodes << std::endl;
+ ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ //preprocessing
+ real* QQ = para->getParH(i)->precursorBC.q27[0];
+ unsigned int sizeQ = para->getParH(i)->precursorBC.numberOfBCnodes;
+ QforBoundaryConditions Q;
+ getPointersToBoundaryConditions(Q, QQ, sizeQ);
+
+ builder->getPrecursorQs(Q.q27, i);
+
+ if (para->getDiffOn()) {
+ throw std::runtime_error("Advection diffusion not implemented for Precursor!");
+ //////////////////////////////////////////////////////////////////////////
+ // para->getParH(i)->TempVel.kTemp = numberOfVelocityNodes;
+ // para->getParD(i)->TempVel.kTemp = numberOfVelocityNodes;
+ // std::cout << "Groesse TempVel.kTemp = " << para->getParH(i)->TempPress.kTemp << std::endl;
+ // std::cout << "getTemperatureInit = " << para->getTemperatureInit() << std::endl;
+ // std::cout << "getTemperatureBC = " << para->getTemperatureBC() << std::endl;
+ // //////////////////////////////////////////////////////////////////////////
+ // cudaMemoryManager->cudaAllocTempVeloBC(i);
+ // //cout << "nach alloc " << std::endl;
+ // //////////////////////////////////////////////////////////////////////////
+ // for (int m = 0; m < numberOfVelocityNodes; m++)
+ // {
+ // para->getParH(i)->TempVel.temp[m] = para->getTemperatureInit();
+ // para->getParH(i)->TempVel.tempPulse[m] = para->getTemperatureBC();
+ // para->getParH(i)->TempVel.velo[m] = para->getVelocity();
+ // para->getParH(i)->TempVel.k[m] = para->getParH(i)->Qinflow.k[m];
+ // }
+ // //////////////////////////////////////////////////////////////////////////
+ // //cout << "vor copy " << std::endl;
+ // cudaMemoryManager->cudaCopyTempVeloBCHD(i);
+ // //cout << "nach copy " << std::endl;
+ //////////////////////////////////////////////////////////////////////////
+ }
+ cudaMemoryManager->cudaCopyPrecursorBC(i);
+ }
+ }
+
+
for (uint i = 0; i < builder->getNumberOfGridLevels(); i++) {
const int numberOfGeometryNodes = builder->getGeometrySize(i);
diff --git a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderGenerator/GridGenerator.h b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderGenerator/GridGenerator.h
index d2f56e1df4ee5658c61b8e8a3e94a820d1a4f2f1..cee661a0c1092e7a3d334018133fd2f963ada991 100644
--- a/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderGenerator/GridGenerator.h
+++ b/src/gpu/VirtualFluids_GPU/DataStructureInitializer/GridReaderGenerator/GridGenerator.h
@@ -40,6 +40,7 @@
#include "LBM/LB.h"
+
class Parameter;
class GridBuilder;
class IndexRearrangementForStreams;
@@ -75,8 +76,10 @@ public:
//! \brief allocates and initialized the sub-grid distances at the boundary conditions
void allocArrays_BoundaryQs() override;
void allocArrays_OffsetScale() override;
- void allocArrays_fluidNodeIndices() override;
- void allocArrays_fluidNodeIndicesBorder() override;
+ void allocArrays_taggedFluidNodes() override;
+
+ void tagFluidNodeIndices(std::vector<uint> taggedFluidNodeIndices, CollisionTemplate tag, uint level) override;
+ void sortFluidNodeTags() override;
virtual void setDimensions() override;
virtual void setBoundingBox() override;
diff --git a/src/gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.cpp b/src/gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.cpp
index bff054eb174a0f5fa34119deedde6f1c9733d83c..b1c398638cff1ec1b6d52f59f8e773183e270331 100644
--- a/src/gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.cpp
+++ b/src/gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.cpp
@@ -35,6 +35,11 @@ void BoundaryConditionFactory::setStressBoundaryCondition(const StressBC boundar
this->stressBoundaryCondition = boundaryConditionType;
}
+void BoundaryConditionFactory::setPrecursorBoundaryCondition(const PrecursorBC boundaryConditionType)
+{
+ this->precursorBoundaryCondition = boundaryConditionType;
+}
+
boundaryCondition BoundaryConditionFactory::getVelocityBoundaryConditionPost(bool isGeometryBC) const
{
const VelocityBC &boundaryCondition =
@@ -132,6 +137,22 @@ boundaryCondition BoundaryConditionFactory::getPressureBoundaryConditionPre() co
case PressureBC::OutflowNonReflective:
return QPressNoRhoDev27;
break;
+ case PressureBC::OutflowNonReflectivePressureCorrection:
+ return QPressZeroRhoOutflowDev27;
+ default:
+ return nullptr;
+ }
+}
+
+precursorBoundaryConditionFunc BoundaryConditionFactory::getPrecursorBoundaryConditionPost() const
+{
+ switch (this->precursorBoundaryCondition) {
+ case PrecursorBC::VelocityPrecursor:
+ return QPrecursorDevCompZeroPress;
+ break;
+ case PrecursorBC::DistributionsPrecursor:
+ return PrecursorDevDistributions;
+ break;
default:
return nullptr;
}
diff --git a/src/gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.h b/src/gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.h
index 9d6872c4847be72dff4be7137b774c8082e39e34..c6877cbfeffe5b32c0c2d336e46b02d68cd946a3 100644
--- a/src/gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.h
+++ b/src/gpu/VirtualFluids_GPU/Factories/BoundaryConditionFactory.h
@@ -42,11 +42,13 @@
#include "Parameter/Parameter.h"
#include "gpu/GridGenerator/grid/BoundaryConditions/Side.h"
+
struct LBMSimulationParameter;
class Parameter;
using boundaryCondition = std::function<void(LBMSimulationParameter *, QforBoundaryConditions *)>;
using boundaryConditionWithParameter = std::function<void(Parameter *, QforBoundaryConditions *, const int level)>;
+using precursorBoundaryConditionFunc = std::function<void(LBMSimulationParameter *, QforPrecursorBoundaryConditions *, real timeRatio, real velocityRatio)>;
class BoundaryConditionFactory
{
@@ -109,6 +111,8 @@ public:
PressureNonEquilibriumCompressible,
//! - OutflowNonReflective = outflow boundary condition, should be combined with VelocityAndPressureCompressible
OutflowNonReflective,
+ //! - OutflowNonreflectivePressureCorrection = like OutflowNonReflective, but also reduces pressure overshoot
+ OutflowNonReflectivePressureCorrection,
//! - NotSpecified = the user did not set a boundary condition
NotSpecified
};
@@ -128,11 +132,21 @@ public:
// enum class OutflowBoundaryCondition {}; // TODO:
// https://git.rz.tu-bs.de/m.schoenherr/VirtualFluids_dev/-/issues/16
+ enum class PrecursorBC {
+ //! - VelocityPrecursor
+ VelocityPrecursor,
+ //! - DisitributionsPrecursor
+ DistributionsPrecursor,
+ //! - NotSpecified = the user did not set a boundary condition
+ NotSpecified
+ };
+
void setVelocityBoundaryCondition(const BoundaryConditionFactory::VelocityBC boundaryConditionType);
void setNoSlipBoundaryCondition(const BoundaryConditionFactory::NoSlipBC boundaryConditionType);
void setSlipBoundaryCondition(const BoundaryConditionFactory::SlipBC boundaryConditionType);
void setPressureBoundaryCondition(const BoundaryConditionFactory::PressureBC boundaryConditionType);
void setStressBoundaryCondition(const BoundaryConditionFactory::StressBC boundaryConditionType);
+ void setPrecursorBoundaryCondition(const BoundaryConditionFactory::PrecursorBC boundaryConditionType);
//! \brief set a boundary condition for the geometry
//! param boundaryConditionType: a velocity, no-slip or slip boundary condition
//! \details suggestions for boundaryConditionType:
@@ -152,6 +166,8 @@ public:
[[nodiscard]] boundaryCondition getSlipBoundaryConditionPost(bool isGeometryBC = false) const;
[[nodiscard]] boundaryCondition getPressureBoundaryConditionPre() const;
[[nodiscard]] boundaryCondition getGeometryBoundaryConditionPost() const;
+ [[nodiscard]] precursorBoundaryConditionFunc getPrecursorBoundaryConditionPost() const;
+
[[nodiscard]] boundaryConditionWithParameter getStressBoundaryConditionPost() const;
@@ -162,6 +178,7 @@ private:
PressureBC pressureBoundaryCondition = PressureBC::NotSpecified;
std::variant<VelocityBC, NoSlipBC, SlipBC> geometryBoundaryCondition = NoSlipBC::NoSlipImplicitBounceBack;
StressBC stressBoundaryCondition = StressBC::NotSpecified;
+ PrecursorBC precursorBoundaryCondition = PrecursorBC::NotSpecified;
// OutflowBoundaryConditon outflowBC // TODO: https://git.rz.tu-bs.de/m.schoenherr/VirtualFluids_dev/-/issues/16
};
diff --git a/src/gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.cpp b/src/gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.cpp
index 22192216927f91c33fafc23c54c3fae334abdd34..f5908517491e252eae5078ec5ede4d5ba452bda7 100644
--- a/src/gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.cpp
+++ b/src/gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.cpp
@@ -5,14 +5,17 @@
#include <math.h>
#include <Parameter/Parameter.h>
+
#include "Parameter/CudaStreamManager.h"
-#include "PreCollisionInteractor/ActuatorLine.h"
+#include "PreCollisionInteractor/ActuatorFarm.h"
#include "PreCollisionInteractor/Probes/Probe.h"
+#include <PreCollisionInteractor/PrecursorWriter.h>
#include "Calculation/PorousMedia.h"
#include "lbm/constants/NumericConstants.h"
+
void CudaMemoryManager::cudaAllocFull(int lev)
{
checkCudaErrors( cudaMallocHost((void**) &(parameter->getParH(lev)->geo ), parameter->getParH(lev)->mem_size_int ));
@@ -239,6 +242,7 @@ void CudaMemoryManager::cudaCopyVeloBC(int lev)
checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->velocityBC.deltaVz, parameter->getParH(lev)->velocityBC.deltaVz, mem_size_inflow_Q_q, cudaMemcpyHostToDevice));
}
+
void CudaMemoryManager::cudaFreeVeloBC(int lev)
{
checkCudaErrors( cudaFreeHost(parameter->getParH(lev)->velocityBC.q27[0] ));
@@ -524,24 +528,24 @@ void CudaMemoryManager::cudaCopyProcessNeighborXIndex(int lev, unsigned int proc
cudaMemcpyHostToDevice));
}
void CudaMemoryManager::cudaCopyProcessNeighborXFsHD(int lev, unsigned int processNeighbor,
- const unsigned int &memsizeFsRecv, int streamIndex)
+ const unsigned int &memsizeFsRecv)
{
- if (streamIndex == -1)
+ if (!parameter->getStreamManager()->streamIsRegistered(CudaStreamIndex::SubDomainBorder))
checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->recvProcessNeighborX[processNeighbor].f[0],
parameter->getParH(lev)->recvProcessNeighborX[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsRecv,
cudaMemcpyHostToDevice));
else
checkCudaErrors( cudaMemcpyAsync(parameter->getParD(lev)->recvProcessNeighborX[processNeighbor].f[0],
- parameter->getParH(lev)->recvProcessNeighborX[processNeighbor].f[0],
- parameter->getD3Qxx() * memsizeFsRecv,
- cudaMemcpyHostToDevice,
- parameter->getStreamManager()->getStream(streamIndex)));
+ parameter->getParH(lev)->recvProcessNeighborX[processNeighbor].f[0],
+ parameter->getD3Qxx() * memsizeFsRecv,
+ cudaMemcpyHostToDevice,
+ parameter->getStreamManager()->getStream(CudaStreamIndex::SubDomainBorder)));
}
void CudaMemoryManager::cudaCopyProcessNeighborXFsDH(int lev, unsigned int processNeighbor,
- const unsigned int &memsizeFsSend, int streamIndex)
-{
- if (streamIndex == -1)
+ const unsigned int &memsizeFsSend)
+{
+ if (!parameter->getStreamManager()->streamIsRegistered(CudaStreamIndex::SubDomainBorder))
checkCudaErrors( cudaMemcpy(parameter->getParH(lev)->sendProcessNeighborX[processNeighbor].f[0],
parameter->getParD(lev)->sendProcessNeighborX[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsSend,
@@ -551,7 +555,7 @@ void CudaMemoryManager::cudaCopyProcessNeighborXFsDH(int lev, unsigned int proce
parameter->getParD(lev)->sendProcessNeighborX[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsSend,
cudaMemcpyDeviceToHost,
- parameter->getStreamManager()->getStream(streamIndex)));
+ parameter->getStreamManager()->getStream(CudaStreamIndex::SubDomainBorder)));
}
void CudaMemoryManager::cudaFreeProcessNeighborX(int lev, unsigned int processNeighbor)
{
@@ -594,35 +598,33 @@ void CudaMemoryManager::cudaCopyProcessNeighborYIndex(int lev, unsigned int proc
parameter->getParH(lev)->recvProcessNeighborY[processNeighbor].memsizeIndex,
cudaMemcpyHostToDevice));
}
-void CudaMemoryManager::cudaCopyProcessNeighborYFsHD(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsRecv,
- int streamIndex)
+void CudaMemoryManager::cudaCopyProcessNeighborYFsHD(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsRecv)
{
- if (streamIndex == -1)
+ if (!parameter->getStreamManager()->streamIsRegistered(CudaStreamIndex::SubDomainBorder))
checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->recvProcessNeighborY[processNeighbor].f[0],
parameter->getParH(lev)->recvProcessNeighborY[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsRecv,
cudaMemcpyHostToDevice));
else
- checkCudaErrors(cudaMemcpyAsync(parameter->getParD(lev)->recvProcessNeighborY[processNeighbor].f[0],
- parameter->getParH(lev)->recvProcessNeighborY[processNeighbor].f[0],
- parameter->getD3Qxx() * memsizeFsRecv,
- cudaMemcpyHostToDevice,
- parameter->getStreamManager()->getStream(streamIndex)));
+ checkCudaErrors( cudaMemcpyAsync(parameter->getParD(lev)->recvProcessNeighborY[processNeighbor].f[0],
+ parameter->getParH(lev)->recvProcessNeighborY[processNeighbor].f[0],
+ parameter->getD3Qxx() * memsizeFsRecv,
+ cudaMemcpyHostToDevice,
+ parameter->getStreamManager()->getStream(CudaStreamIndex::SubDomainBorder)));
}
-void CudaMemoryManager::cudaCopyProcessNeighborYFsDH(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsSend,
- int streamIndex)
+void CudaMemoryManager::cudaCopyProcessNeighborYFsDH(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsSend)
{
- if (streamIndex == -1)
+ if (!parameter->getStreamManager()->streamIsRegistered(CudaStreamIndex::SubDomainBorder))
checkCudaErrors( cudaMemcpy(parameter->getParH(lev)->sendProcessNeighborY[processNeighbor].f[0],
parameter->getParD(lev)->sendProcessNeighborY[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsSend,
cudaMemcpyDeviceToHost));
else
- checkCudaErrors(
- cudaMemcpyAsync(parameter->getParH(lev)->sendProcessNeighborY[processNeighbor].f[0],
- parameter->getParD(lev)->sendProcessNeighborY[processNeighbor].f[0],
- parameter->getD3Qxx() * memsizeFsSend,
- cudaMemcpyDeviceToHost, parameter->getStreamManager()->getStream(streamIndex)));
+ checkCudaErrors( cudaMemcpyAsync(parameter->getParH(lev)->sendProcessNeighborY[processNeighbor].f[0],
+ parameter->getParD(lev)->sendProcessNeighborY[processNeighbor].f[0],
+ parameter->getD3Qxx() * memsizeFsSend,
+ cudaMemcpyDeviceToHost,
+ parameter->getStreamManager()->getStream(CudaStreamIndex::SubDomainBorder)));
}
void CudaMemoryManager::cudaFreeProcessNeighborY(int lev, unsigned int processNeighbor)
{
@@ -666,9 +668,9 @@ void CudaMemoryManager::cudaCopyProcessNeighborZIndex(int lev, unsigned int proc
cudaMemcpyHostToDevice));
}
void CudaMemoryManager::cudaCopyProcessNeighborZFsHD(int lev, unsigned int processNeighbor,
- const unsigned int &memsizeFsRecv, int streamIndex)
+ const unsigned int &memsizeFsRecv)
{
- if (streamIndex == -1)
+ if (!parameter->getStreamManager()->streamIsRegistered(CudaStreamIndex::SubDomainBorder))
checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->recvProcessNeighborZ[processNeighbor].f[0],
parameter->getParH(lev)->recvProcessNeighborZ[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsRecv,
@@ -678,12 +680,12 @@ void CudaMemoryManager::cudaCopyProcessNeighborZFsHD(int lev, unsigned int proce
parameter->getParH(lev)->recvProcessNeighborZ[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsRecv,
cudaMemcpyHostToDevice,
- parameter->getStreamManager()->getStream(streamIndex)));
+ parameter->getStreamManager()->getStream(CudaStreamIndex::SubDomainBorder)));
}
void CudaMemoryManager::cudaCopyProcessNeighborZFsDH(int lev, unsigned int processNeighbor,
- const unsigned int &memsizeFsSend, int streamIndex)
+ const unsigned int &memsizeFsSend)
{
- if (streamIndex == -1)
+ if (!parameter->getStreamManager()->streamIsRegistered(CudaStreamIndex::SubDomainBorder))
checkCudaErrors( cudaMemcpy(parameter->getParH(lev)->sendProcessNeighborZ[processNeighbor].f[0],
parameter->getParD(lev)->sendProcessNeighborZ[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsSend,
@@ -693,7 +695,7 @@ void CudaMemoryManager::cudaCopyProcessNeighborZFsDH(int lev, unsigned int proce
parameter->getParD(lev)->sendProcessNeighborZ[processNeighbor].f[0],
parameter->getD3Qxx() * memsizeFsSend,
cudaMemcpyDeviceToHost,
- parameter->getStreamManager()->getStream(streamIndex)));
+ parameter->getStreamManager()->getStream(CudaStreamIndex::SubDomainBorder)));
}
void CudaMemoryManager::cudaFreeProcessNeighborZ(int lev, unsigned int processNeighbor)
{
@@ -1655,6 +1657,133 @@ void CudaMemoryManager::cudaFreeWallModel(int lev, bool hasWallModelMonitor)
}
}
+
+//Precursor BC
+void CudaMemoryManager::cudaAllocPrecursorBC(int lev)
+{
+ uint memSizeQInt = parameter->getParH(lev)->precursorBC.numberOfBCnodes*sizeof(int);
+ uint memSizeQUint = parameter->getParH(lev)->precursorBC.numberOfBCnodes*sizeof(uint);
+ uint memSizeQReal = parameter->getParH(lev)->precursorBC.numberOfBCnodes*sizeof(real);
+
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.k, memSizeQInt));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.q27[0], parameter->getD3Qxx()*memSizeQReal));
+
+
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.planeNeighbor0PP, memSizeQUint));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.planeNeighbor0PM, memSizeQUint));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.planeNeighbor0MP, memSizeQUint));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.planeNeighbor0MM, memSizeQUint));
+
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.weights0PP, memSizeQReal));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.weights0PM, memSizeQReal));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.weights0MP, memSizeQReal));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.weights0MM, memSizeQReal));
+
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.k, memSizeQInt));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.q27[0], parameter->getD3Qxx()*memSizeQReal));
+
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.planeNeighbor0PP, memSizeQUint));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.planeNeighbor0PM, memSizeQUint));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.planeNeighbor0MP, memSizeQUint));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.planeNeighbor0MM, memSizeQUint));
+
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.weights0PP, memSizeQReal));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.weights0PM, memSizeQReal));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.weights0MP, memSizeQReal));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.weights0MM, memSizeQReal));
+
+ real memSize = memSizeQInt+4*memSizeQUint+(4+parameter->getD3Qxx())*memSizeQReal;
+ setMemsizeGPU(memSize, false);
+
+}
+
+
+void CudaMemoryManager::cudaAllocPrecursorData(int lev)
+{
+ size_t size = parameter->getParH(lev)->precursorBC.numberOfPrecursorNodes*sizeof(real)*parameter->getParH(lev)->precursorBC.numberOfQuantities;
+
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.last, size));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.current, size));
+ checkCudaErrors( cudaMallocHost((void**) ¶meter->getParH(lev)->precursorBC.next, size));
+
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.last, size));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.current, size));
+ checkCudaErrors( cudaMalloc((void**) ¶meter->getParD(lev)->precursorBC.next, size));
+ setMemsizeGPU(3*size, false);
+}
+
+
+void CudaMemoryManager::cudaCopyPrecursorBC(int lev)
+{
+ uint memSizeQInt = parameter->getParH(lev)->precursorBC.numberOfBCnodes*sizeof(int);
+ uint memSizeQUint = parameter->getParH(lev)->precursorBC.numberOfBCnodes*sizeof(uint);
+ uint memSizeQReal = parameter->getParH(lev)->precursorBC.numberOfBCnodes*sizeof(real);
+
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.k, parameter->getParH(lev)->precursorBC.k, memSizeQInt, cudaMemcpyHostToDevice));
+
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.q27[0], parameter->getParH(lev)->precursorBC.q27[0], memSizeQReal*parameter->getD3Qxx(), cudaMemcpyHostToDevice));
+
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.planeNeighbor0PP, parameter->getParH(lev)->precursorBC.planeNeighbor0PP, memSizeQUint, cudaMemcpyHostToDevice));
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.planeNeighbor0PM, parameter->getParH(lev)->precursorBC.planeNeighbor0PM, memSizeQUint, cudaMemcpyHostToDevice));
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.planeNeighbor0MP, parameter->getParH(lev)->precursorBC.planeNeighbor0MP, memSizeQUint, cudaMemcpyHostToDevice));
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.planeNeighbor0MM, parameter->getParH(lev)->precursorBC.planeNeighbor0MM, memSizeQUint, cudaMemcpyHostToDevice));
+
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.weights0PP, parameter->getParH(lev)->precursorBC.weights0PP, memSizeQReal, cudaMemcpyHostToDevice));
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.weights0PM, parameter->getParH(lev)->precursorBC.weights0PM, memSizeQReal, cudaMemcpyHostToDevice));
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.weights0MP, parameter->getParH(lev)->precursorBC.weights0MP, memSizeQReal, cudaMemcpyHostToDevice));
+ checkCudaErrors( cudaMemcpy(parameter->getParD(lev)->precursorBC.weights0MM, parameter->getParH(lev)->precursorBC.weights0MM, memSizeQReal, cudaMemcpyHostToDevice));
+}
+void CudaMemoryManager::cudaCopyPrecursorData(int lev)
+{
+ auto prec = ¶meter->getParH(lev)->precursorBC;
+ auto precStream = parameter->getStreamManager()->getStream(CudaStreamIndex::Precursor);
+ size_t memSize = prec->numberOfPrecursorNodes*sizeof(real)*prec->numberOfQuantities;
+ checkCudaErrors( cudaStreamSynchronize(precStream) );
+ checkCudaErrors( cudaMemcpyAsync(parameter->getParD(lev)->precursorBC.next, prec->next, memSize, cudaMemcpyHostToDevice, precStream) );
+}
+
+
+void CudaMemoryManager::cudaFreePrecursorBC(int lev)
+{
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.k));
+
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.q27[0]));
+
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.planeNeighbor0PP));
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.planeNeighbor0PM));
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.planeNeighbor0MP));
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.planeNeighbor0MM));
+
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.weights0PP));
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.weights0PM));
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.weights0MP));
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.weights0MM));
+
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.k));
+
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.q27[0]));
+
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.planeNeighbor0PP));
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.planeNeighbor0PM));
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.planeNeighbor0MP));
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.planeNeighbor0MM));
+
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.weights0PP));
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.weights0PM));
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.weights0MP));
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.weights0MM));
+}
+
+void CudaMemoryManager::cudaFreePrecursorData(int lev)
+{
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.last));
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.current));
+ checkCudaErrors( cudaFreeHost( parameter->getParH(lev)->precursorBC.next));
+
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.last));
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.current));
+ checkCudaErrors( cudaFree( parameter->getParD(lev)->precursorBC.next));
+}
//Test roundoff error
void CudaMemoryManager::cudaAllocTestRE(int lev, unsigned int size)
{
@@ -2921,231 +3050,297 @@ void CudaMemoryManager::cudaFree2ndOrderDerivitivesIsoTest(int lev)
}
-void CudaMemoryManager::cudaAllocFluidNodeIndices(int lev) {
- uint mem_size_geo_fluid_nodes = sizeof(uint) * parameter->getParH(lev)->numberOfFluidNodes;
+void CudaMemoryManager::cudaAllocTaggedFluidNodeIndices(CollisionTemplate tag, int lev) {
+ uint mem_size_tagged_fluid_nodes = sizeof(uint) * parameter->getParH(lev)->numberOfTaggedFluidNodes[tag];
// Host
- checkCudaErrors(cudaMallocHost((void **)&(parameter->getParH(lev)->fluidNodeIndices), mem_size_geo_fluid_nodes));
+ checkCudaErrors(cudaMallocHost((void **)&(parameter->getParH(lev)->taggedFluidNodeIndices[tag]), mem_size_tagged_fluid_nodes));
// Device
- checkCudaErrors(cudaMalloc((void **)&(parameter->getParD(lev)->fluidNodeIndices), mem_size_geo_fluid_nodes));
+ checkCudaErrors(cudaMalloc((void **)&(parameter->getParD(lev)->taggedFluidNodeIndices[tag]), mem_size_tagged_fluid_nodes));
//////////////////////////////////////////////////////////////////////////
- setMemsizeGPU((double)mem_size_geo_fluid_nodes, false);
+ setMemsizeGPU((double)mem_size_tagged_fluid_nodes, false);
}
-void CudaMemoryManager::cudaCopyFluidNodeIndices(int lev) {
- uint mem_size_geo_fluid_nodes = sizeof(uint) * parameter->getParH(lev)->numberOfFluidNodes;
- checkCudaErrors(cudaMemcpy(parameter->getParD(lev)->fluidNodeIndices,
- parameter->getParH(lev)->fluidNodeIndices,
- mem_size_geo_fluid_nodes, cudaMemcpyHostToDevice));
+void CudaMemoryManager::cudaCopyTaggedFluidNodeIndices(CollisionTemplate tag, int lev) {
+ uint mem_size_tagged_fluid_nodes = sizeof(uint) * parameter->getParH(lev)->numberOfTaggedFluidNodes[tag];
+ checkCudaErrors(cudaMemcpy(parameter->getParD(lev)->taggedFluidNodeIndices[tag],
+ parameter->getParH(lev)->taggedFluidNodeIndices[tag],
+ mem_size_tagged_fluid_nodes, cudaMemcpyHostToDevice));
}
-void CudaMemoryManager::cudaFreeFluidNodeIndices(int lev) {
- checkCudaErrors(cudaFreeHost(parameter->getParH(lev)->fluidNodeIndices));
+void CudaMemoryManager::cudaFreeTaggedFluidNodeIndices(CollisionTemplate tag, int lev) {
+ checkCudaErrors(cudaFreeHost(parameter->getParH(lev)->taggedFluidNodeIndices[tag]));
}
-void CudaMemoryManager::cudaAllocFluidNodeIndicesBorder(int lev) {
- uint mem_size_fluid_nodes_border = sizeof(uint) * parameter->getParH(lev)->numberOfFluidNodesBorder;
- // Host
- checkCudaErrors(
- cudaMallocHost((void **)&(parameter->getParH(lev)->fluidNodeIndicesBorder), mem_size_fluid_nodes_border));
- // Device
- checkCudaErrors(
- cudaMalloc((void **)&(parameter->getParD(lev)->fluidNodeIndicesBorder), mem_size_fluid_nodes_border));
- //////////////////////////////////////////////////////////////////////////
- setMemsizeGPU((double)mem_size_fluid_nodes_border, false);
-}
+////////////////////////////////////////////////////////////////////////////////////
+// ActuatorFarm
+///////////////////////////////////////////////////////////////////////////////
+void CudaMemoryManager::cudaAllocBladeGeometries(ActuatorFarm* actuatorFarm)
+{
+ uint sizeRealTurbine = sizeof(real)*actuatorFarm->getNumberOfTurbines();
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeRadiiH, sizeRealTurbine*actuatorFarm->getNumberOfNodesPerBlade()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->diametersH, sizeRealTurbine) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->turbinePosXH, sizeRealTurbine) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->turbinePosYH, sizeRealTurbine) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->turbinePosZH, sizeRealTurbine) );
+
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeRadiiD, sizeRealTurbine*actuatorFarm->getNumberOfNodesPerBlade()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->diametersD, sizeRealTurbine) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->turbinePosXD, sizeRealTurbine) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->turbinePosYD, sizeRealTurbine) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->turbinePosZD, sizeRealTurbine) );
+ setMemsizeGPU(sizeof(real)*(actuatorFarm->getNumberOfNodesPerBlade()+4)*actuatorFarm->getNumberOfTurbines(), false);
-void CudaMemoryManager::cudaCopyFluidNodeIndicesBorder(int lev) {
- uint mem_size_fluid_nodes_border = sizeof(uint) * parameter->getParH(lev)->numberOfFluidNodesBorder;
- checkCudaErrors(cudaMemcpy(parameter->getParD(lev)->fluidNodeIndicesBorder,
- parameter->getParH(lev)->fluidNodeIndicesBorder,
- mem_size_fluid_nodes_border, cudaMemcpyHostToDevice));
}
+void CudaMemoryManager::cudaCopyBladeGeometriesHtoD(ActuatorFarm* actuatorFarm)
+{
+ uint sizeRealTurbine = sizeof(real)*actuatorFarm->getNumberOfTurbines();
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeRadiiD, actuatorFarm->bladeRadiiH, sizeRealTurbine*actuatorFarm->getNumberOfNodesPerBlade(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->diametersD, actuatorFarm->diametersH, sizeRealTurbine, cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->turbinePosXD, actuatorFarm->turbinePosXH, sizeRealTurbine, cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->turbinePosYD, actuatorFarm->turbinePosYH, sizeRealTurbine, cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->turbinePosZD, actuatorFarm->turbinePosZH, sizeRealTurbine, cudaMemcpyHostToDevice) );
-void CudaMemoryManager::cudaFreeFluidNodeIndicesBorder(int lev) {
- checkCudaErrors(cudaFreeHost(parameter->getParH(lev)->fluidNodeIndicesBorder));
}
+void CudaMemoryManager::cudaCopyBladeGeometriesDtoH(ActuatorFarm* actuatorFarm)
+{
+ uint sizeRealTurbine = sizeof(real)*actuatorFarm->getNumberOfTurbines();
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeRadiiH, actuatorFarm->bladeRadiiD, sizeRealTurbine*actuatorFarm->getNumberOfNodesPerBlade(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->diametersH, actuatorFarm->diametersD, sizeRealTurbine, cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->turbinePosXH, actuatorFarm->turbinePosXD, sizeRealTurbine, cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->turbinePosYH, actuatorFarm->turbinePosYD, sizeRealTurbine, cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->turbinePosZH, actuatorFarm->turbinePosZD, sizeRealTurbine, cudaMemcpyDeviceToHost) );
-////////////////////////////////////////////////////////////////////////////////////
-// ActuatorLine
-///////////////////////////////////////////////////////////////////////////////
+}
+void CudaMemoryManager::cudaFreeBladeGeometries(ActuatorFarm* actuatorFarm)
+{
+ checkCudaErrors( cudaFree(actuatorFarm->bladeRadiiD) );
+ checkCudaErrors( cudaFree(actuatorFarm->diametersD) );
+ checkCudaErrors( cudaFree(actuatorFarm->turbinePosXD) );
+ checkCudaErrors( cudaFree(actuatorFarm->turbinePosYD) );
+ checkCudaErrors( cudaFree(actuatorFarm->turbinePosZD) );
+
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeRadiiH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->diametersH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->turbinePosXH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->turbinePosYH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->turbinePosZH) );
+}
-void CudaMemoryManager::cudaAllocBladeRadii(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaAllocBladeOrientations(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeRadiiH, sizeof(real)*actuatorLine->getNBladeNodes()) );
+ uint sizeRealTurbine = sizeof(real)*actuatorFarm->getNumberOfTurbines();
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->omegasH, sizeRealTurbine) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->azimuthsH, sizeRealTurbine) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->yawsH, sizeRealTurbine) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeRadiiD, sizeof(real)*actuatorLine->getNBladeNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->omegasD, sizeRealTurbine) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->azimuthsD, sizeRealTurbine) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->yawsD, sizeRealTurbine) );
- setMemsizeGPU(sizeof(real)*actuatorLine->getNBladeNodes(), false);
-}
+ setMemsizeGPU(3*sizeRealTurbine, false);
-void CudaMemoryManager::cudaCopyBladeRadiiHtoD(ActuatorLine* actuatorLine)
-{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeRadiiD, actuatorLine->bladeRadiiH, sizeof(real)*actuatorLine->getNBladeNodes(), cudaMemcpyHostToDevice) );
}
+void CudaMemoryManager::cudaCopyBladeOrientationsHtoD(ActuatorFarm* actuatorFarm)
+{
+ uint sizeRealTurbine = sizeof(real)*actuatorFarm->getNumberOfTurbines();
+ checkCudaErrors( cudaMemcpy(actuatorFarm->omegasD, actuatorFarm->omegasH, sizeRealTurbine, cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->azimuthsD, actuatorFarm->azimuthsH, sizeRealTurbine, cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->yawsD, actuatorFarm->yawsH, sizeRealTurbine, cudaMemcpyHostToDevice) );
-void CudaMemoryManager::cudaCopyBladeRadiiDtoH(ActuatorLine* actuatorLine)
+}
+void CudaMemoryManager::cudaCopyBladeOrientationsDtoH(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeRadiiH, actuatorLine->bladeRadiiD, sizeof(real)*actuatorLine->getNBladeNodes(), cudaMemcpyDeviceToHost) );
+ uint sizeRealTurbine = sizeof(real)*actuatorFarm->getNumberOfTurbines();
+ checkCudaErrors( cudaMemcpy(actuatorFarm->omegasH, actuatorFarm->omegasD, sizeRealTurbine, cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->azimuthsH, actuatorFarm->azimuthsD, sizeRealTurbine, cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->yawsH, actuatorFarm->yawsD, sizeRealTurbine, cudaMemcpyDeviceToHost) );
}
-
-void CudaMemoryManager::cudaFreeBladeRadii(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaFreeBladeOrientations(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaFree(actuatorLine->bladeRadiiD) );
+ checkCudaErrors( cudaFree((void**) &actuatorFarm->omegasD) );
+ checkCudaErrors( cudaFree((void**) &actuatorFarm->azimuthsD) );
+ checkCudaErrors( cudaFree((void**) &actuatorFarm->yawsD) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeRadiiH) );
+ checkCudaErrors( cudaFreeHost((void**) &actuatorFarm->omegasH) );
+ checkCudaErrors( cudaFreeHost((void**) &actuatorFarm->azimuthsH) );
+ checkCudaErrors( cudaFreeHost((void**) &actuatorFarm->yawsH) );
}
-void CudaMemoryManager::cudaAllocBladeCoords(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaAllocBladeCoords(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeCoordsXH, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeCoordsYH, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeCoordsZH, sizeof(real)*actuatorLine->getNNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeCoordsXH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeCoordsYH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeCoordsZH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeCoordsXD, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeCoordsYD, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeCoordsZD, sizeof(real)*actuatorLine->getNNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeCoordsXDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeCoordsYDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeCoordsZDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeCoordsXDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeCoordsYDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeCoordsZDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
- setMemsizeGPU(3.f*actuatorLine->getNNodes(), false);
+ setMemsizeGPU(6.f*actuatorFarm->getNumberOfNodes(), false);
}
-void CudaMemoryManager::cudaCopyBladeCoordsHtoD(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaCopyBladeCoordsHtoD(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeCoordsXD, actuatorLine->bladeCoordsXH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeCoordsYD, actuatorLine->bladeCoordsYH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeCoordsZD, actuatorLine->bladeCoordsZH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeCoordsXDCurrentTimestep, actuatorFarm->bladeCoordsXH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeCoordsYDCurrentTimestep, actuatorFarm->bladeCoordsYH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeCoordsZDCurrentTimestep, actuatorFarm->bladeCoordsZH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
}
-void CudaMemoryManager::cudaCopyBladeCoordsDtoH(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaCopyBladeCoordsDtoH(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeCoordsXH, actuatorLine->bladeCoordsXD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeCoordsYH, actuatorLine->bladeCoordsYD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeCoordsZH, actuatorLine->bladeCoordsZD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeCoordsXH, actuatorFarm->bladeCoordsXDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeCoordsYH, actuatorFarm->bladeCoordsYDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeCoordsZH, actuatorFarm->bladeCoordsZDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
}
-void CudaMemoryManager::cudaFreeBladeCoords(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaFreeBladeCoords(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaFree(actuatorLine->bladeCoordsXD) );
- checkCudaErrors( cudaFree(actuatorLine->bladeCoordsYD) );
- checkCudaErrors( cudaFree(actuatorLine->bladeCoordsZD) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeCoordsXDCurrentTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeCoordsYDCurrentTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeCoordsZDCurrentTimestep) );
+
+ checkCudaErrors( cudaFree(actuatorFarm->bladeCoordsXDPreviousTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeCoordsYDPreviousTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeCoordsZDPreviousTimestep) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeCoordsXH) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeCoordsYH) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeCoordsZH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeCoordsXH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeCoordsYH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeCoordsZH) );
}
-void CudaMemoryManager::cudaAllocBladeIndices(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaAllocBladeIndices(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeIndicesH, sizeof(uint)*actuatorLine->getNNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeIndicesH, sizeof(uint)*actuatorFarm->getNumberOfNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeIndicesD, sizeof(uint)*actuatorLine->getNNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeIndicesD, sizeof(uint)*actuatorFarm->getNumberOfNodes()) );
- setMemsizeGPU(sizeof(uint)*actuatorLine->getNNodes(), false);
+ setMemsizeGPU(sizeof(uint)*actuatorFarm->getNumberOfNodes(), false);
}
-void CudaMemoryManager::cudaCopyBladeIndicesHtoD(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaCopyBladeIndicesHtoD(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeIndicesD, actuatorLine->bladeIndicesH, sizeof(uint)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeIndicesD, actuatorFarm->bladeIndicesH, sizeof(uint)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
}
-void CudaMemoryManager::cudaFreeBladeIndices(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaFreeBladeIndices(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaFree(actuatorLine->bladeIndicesD) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeIndicesD) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeIndicesH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeIndicesH) );
}
-void CudaMemoryManager::cudaAllocBladeVelocities(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaAllocBladeVelocities(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeVelocitiesXH, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeVelocitiesYH, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeVelocitiesZH, sizeof(real)*actuatorLine->getNNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeVelocitiesXH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeVelocitiesYH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeVelocitiesZH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeVelocitiesXD, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeVelocitiesYD, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeVelocitiesZD, sizeof(real)*actuatorLine->getNNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeVelocitiesXDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeVelocitiesYDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeVelocitiesZDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
- setMemsizeGPU(3.*sizeof(real)*actuatorLine->getNNodes(), false);
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeVelocitiesXDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeVelocitiesYDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeVelocitiesZDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+
+ setMemsizeGPU(3.*sizeof(real)*actuatorFarm->getNumberOfNodes(), false);
}
-void CudaMemoryManager::cudaCopyBladeVelocitiesHtoD(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaCopyBladeVelocitiesHtoD(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeVelocitiesXD, actuatorLine->bladeVelocitiesXH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeVelocitiesYD, actuatorLine->bladeVelocitiesYH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeVelocitiesZD, actuatorLine->bladeVelocitiesZH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeVelocitiesXDCurrentTimestep, actuatorFarm->bladeVelocitiesXH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeVelocitiesYDCurrentTimestep, actuatorFarm->bladeVelocitiesYH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeVelocitiesZDCurrentTimestep, actuatorFarm->bladeVelocitiesZH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
}
-void CudaMemoryManager::cudaCopyBladeVelocitiesDtoH(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaCopyBladeVelocitiesDtoH(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeVelocitiesXH, actuatorLine->bladeVelocitiesXD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeVelocitiesYH, actuatorLine->bladeVelocitiesYD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeVelocitiesZH, actuatorLine->bladeVelocitiesZD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeVelocitiesXH, actuatorFarm->bladeVelocitiesXDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeVelocitiesYH, actuatorFarm->bladeVelocitiesYDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeVelocitiesZH, actuatorFarm->bladeVelocitiesZDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
}
-void CudaMemoryManager::cudaFreeBladeVelocities(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaFreeBladeVelocities(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaFree(actuatorLine->bladeVelocitiesXD) );
- checkCudaErrors( cudaFree(actuatorLine->bladeVelocitiesYD) );
- checkCudaErrors( cudaFree(actuatorLine->bladeVelocitiesZD) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeVelocitiesXDCurrentTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeVelocitiesYDCurrentTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeVelocitiesZDCurrentTimestep) );
+
+ checkCudaErrors( cudaFree(actuatorFarm->bladeVelocitiesXDPreviousTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeVelocitiesYDPreviousTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeVelocitiesZDPreviousTimestep) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeVelocitiesXH) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeVelocitiesYH) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeVelocitiesZH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeVelocitiesXH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeVelocitiesYH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeVelocitiesZH) );
}
-void CudaMemoryManager::cudaAllocBladeForces(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaAllocBladeForces(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeForcesXH, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeForcesYH, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMallocHost((void**) &actuatorLine->bladeForcesZH, sizeof(real)*actuatorLine->getNNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeForcesXH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeForcesYH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMallocHost((void**) &actuatorFarm->bladeForcesZH, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeForcesXDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeForcesYDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeForcesZDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeForcesXD, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeForcesYD, sizeof(real)*actuatorLine->getNNodes()) );
- checkCudaErrors( cudaMalloc((void**) &actuatorLine->bladeForcesZD, sizeof(real)*actuatorLine->getNNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeForcesXDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeForcesYDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
+ checkCudaErrors( cudaMalloc((void**) &actuatorFarm->bladeForcesZDPreviousTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes()) );
- setMemsizeGPU(3.*sizeof(real)*actuatorLine->getNNodes(), false);
+ setMemsizeGPU(3.*sizeof(real)*actuatorFarm->getNumberOfNodes(), false);
}
-void CudaMemoryManager::cudaCopyBladeForcesHtoD(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaCopyBladeForcesHtoD(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeForcesXD, actuatorLine->bladeForcesXH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeForcesYD, actuatorLine->bladeForcesYH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeForcesZD, actuatorLine->bladeForcesZH, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeForcesXDCurrentTimestep, actuatorFarm->bladeForcesXH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeForcesYDCurrentTimestep, actuatorFarm->bladeForcesYH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeForcesZDCurrentTimestep, actuatorFarm->bladeForcesZH, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyHostToDevice) );
}
-void CudaMemoryManager::cudaCopyBladeForcesDtoH(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaCopyBladeForcesDtoH(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeForcesXH, actuatorLine->bladeForcesXD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeForcesYH, actuatorLine->bladeForcesYD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
- checkCudaErrors( cudaMemcpy(actuatorLine->bladeForcesZH, actuatorLine->bladeForcesZD, sizeof(real)*actuatorLine->getNNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeForcesXH, actuatorFarm->bladeForcesXDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeForcesYH, actuatorFarm->bladeForcesYDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->bladeForcesZH, actuatorFarm->bladeForcesZDCurrentTimestep, sizeof(real)*actuatorFarm->getNumberOfNodes(), cudaMemcpyDeviceToHost) );
}
-void CudaMemoryManager::cudaFreeBladeForces(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaFreeBladeForces(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaFree(actuatorLine->bladeForcesXD) );
- checkCudaErrors( cudaFree(actuatorLine->bladeForcesYD) );
- checkCudaErrors( cudaFree(actuatorLine->bladeForcesZD) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeForcesXDCurrentTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeForcesYDCurrentTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeForcesZDCurrentTimestep) );
+
+ checkCudaErrors( cudaFree(actuatorFarm->bladeForcesXDPreviousTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeForcesYDPreviousTimestep) );
+ checkCudaErrors( cudaFree(actuatorFarm->bladeForcesZDPreviousTimestep) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeForcesXH) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeForcesYH) );
- checkCudaErrors( cudaFreeHost(actuatorLine->bladeForcesZH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeForcesXH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeForcesYH) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->bladeForcesZH) );
}
-void CudaMemoryManager::cudaAllocSphereIndices(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaAllocSphereIndices(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMallocHost((void**) &(actuatorLine->boundingSphereIndicesH), sizeof(int)*actuatorLine->getNIndices()));
- checkCudaErrors( cudaMalloc((void**) &(actuatorLine->boundingSphereIndicesD), sizeof(int)*actuatorLine->getNIndices()));
- setMemsizeGPU(sizeof(int)*actuatorLine->getNIndices(), false);
+ checkCudaErrors( cudaMallocHost((void**) &(actuatorFarm->boundingSphereIndicesH), sizeof(int)*actuatorFarm->getNumberOfIndices()));
+ checkCudaErrors( cudaMalloc((void**) &(actuatorFarm->boundingSphereIndicesD), sizeof(int)*actuatorFarm->getNumberOfIndices()));
+ setMemsizeGPU(sizeof(int)*actuatorFarm->getNumberOfIndices(), false);
}
-void CudaMemoryManager::cudaCopySphereIndicesHtoD(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaCopySphereIndicesHtoD(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaMemcpy(actuatorLine->boundingSphereIndicesD, actuatorLine->boundingSphereIndicesH, sizeof(int)*actuatorLine->getNIndices(), cudaMemcpyHostToDevice) );
+ checkCudaErrors( cudaMemcpy(actuatorFarm->boundingSphereIndicesD, actuatorFarm->boundingSphereIndicesH, sizeof(int)*actuatorFarm->getNumberOfIndices(), cudaMemcpyHostToDevice) );
}
-void CudaMemoryManager::cudaFreeSphereIndices(ActuatorLine* actuatorLine)
+void CudaMemoryManager::cudaFreeSphereIndices(ActuatorFarm* actuatorFarm)
{
- checkCudaErrors( cudaFreeHost(actuatorLine->boundingSphereIndicesH) );
- checkCudaErrors( cudaFree(actuatorLine->boundingSphereIndicesD) );
+ checkCudaErrors( cudaFreeHost(actuatorFarm->boundingSphereIndicesH) );
+ checkCudaErrors( cudaFree(actuatorFarm->boundingSphereIndicesD) );
}
////////////////////////////////////////////////////////////////////////////////////
@@ -3231,8 +3426,11 @@ void CudaMemoryManager::cudaCopyProbeQuantityArrayHtoD(Probe* probe, int level)
}
void CudaMemoryManager::cudaCopyProbeQuantityArrayDtoH(Probe* probe, int level)
{
- checkCudaErrors( cudaMemcpy(probe->getProbeStruct(level)->quantitiesArrayH, probe->getProbeStruct(level)->quantitiesArrayD, probe->getProbeStruct(level)->nArrays*sizeof(real)*probe->getProbeStruct(level)->nPoints, cudaMemcpyDeviceToHost) );
+ auto probeStruct = probe->getProbeStruct(level);
+
+ checkCudaErrors( cudaMemcpy(probeStruct->quantitiesArrayH, probeStruct->quantitiesArrayD, probeStruct->nArrays*sizeof(real)*probeStruct->nPoints, cudaMemcpyDeviceToHost) );
}
+
void CudaMemoryManager::cudaFreeProbeQuantityArray(Probe* probe, int level)
{
checkCudaErrors( cudaFreeHost(probe->getProbeStruct(level)->quantitiesArrayH) );
@@ -3262,6 +3460,7 @@ void CudaMemoryManager::cudaCopyProbeQuantitiesAndOffsetsDtoH(Probe* probe, int
checkCudaErrors( cudaMemcpy(probe->getProbeStruct(level)->quantitiesH, probe->getProbeStruct(level)->quantitiesD, int(Statistic::LAST)*sizeof(bool), cudaMemcpyDeviceToHost) );
checkCudaErrors( cudaMemcpy(probe->getProbeStruct(level)->arrayOffsetsH, probe->getProbeStruct(level)->arrayOffsetsD, int(Statistic::LAST)*sizeof(int), cudaMemcpyDeviceToHost) );
}
+
void CudaMemoryManager::cudaFreeProbeQuantitiesAndOffsets(Probe* probe, int level)
{
checkCudaErrors( cudaFreeHost(probe->getProbeStruct(level)->quantitiesH) );
@@ -3270,23 +3469,51 @@ void CudaMemoryManager::cudaFreeProbeQuantitiesAndOffsets(Probe* probe, int leve
checkCudaErrors( cudaFree (probe->getProbeStruct(level)->arrayOffsetsD) );
}
+void CudaMemoryManager::cudaAllocPrecursorWriter(PrecursorWriter* writer, int level)
+{
+ auto prec = writer->getPrecursorStruct(level);
+ size_t indSize = prec->numberOfPointsInBC*sizeof(uint);
+ checkCudaErrors( cudaStreamCreate(&prec->stream) );
+ checkCudaErrors( cudaMallocHost((void**) &prec->indicesH, indSize));
+ checkCudaErrors( cudaMalloc((void**) &prec->indicesD, indSize));
+ size_t dataSize = prec->numberOfPointsInBC*sizeof(real)*prec->numberOfQuantities;
+ size_t dataSizeH = dataSize * prec->numberOfTimestepsPerFile;
+
+ checkCudaErrors( cudaMallocHost((void**) &prec->dataH, dataSizeH));
+ checkCudaErrors( cudaMallocHost((void**) &prec->bufferH, dataSizeH));
+ checkCudaErrors( cudaMalloc((void**) &prec->dataD, dataSize));
+ checkCudaErrors( cudaMalloc((void**) &prec->bufferD, dataSize));
+ setMemsizeGPU(indSize+2*dataSize, false);
+}
+void CudaMemoryManager::cudaCopyPrecursorWriterIndicesHtoD(PrecursorWriter* writer, int level)
+{
+ checkCudaErrors( cudaMemcpy(writer->getPrecursorStruct(level)->indicesD, writer->getPrecursorStruct(level)->indicesH, writer->getPrecursorStruct(level)->numberOfPointsInBC*sizeof(uint), cudaMemcpyHostToDevice) );
+}
+void CudaMemoryManager::cudaCopyPrecursorWriterOutputVariablesDtoH(PrecursorWriter* writer, int level)
+{
+ auto prec = writer->getPrecursorStruct(level);
+ int sizeTimestep = prec->numberOfPointsInBC*prec->numberOfQuantities;
+ checkCudaErrors( cudaStreamSynchronize(prec->stream) );
+ checkCudaErrors( cudaMemcpyAsync( &prec->bufferH[prec->numberOfTimestepsBuffered*sizeTimestep], prec->bufferD, sizeof(real)*sizeTimestep, cudaMemcpyDeviceToHost, prec->stream));
+}
+void CudaMemoryManager::cudaFreePrecursorWriter(PrecursorWriter* writer, int level)
+{
+ checkCudaErrors( cudaFreeHost(writer->getPrecursorStruct(level)->indicesH));
+ checkCudaErrors( cudaFree(writer->getPrecursorStruct(level)->indicesD));
-
-
-
-
-
-
-
-
+ checkCudaErrors( cudaFreeHost(writer->getPrecursorStruct(level)->dataH));
+ checkCudaErrors( cudaFreeHost(writer->getPrecursorStruct(level)->bufferH));
+ checkCudaErrors( cudaFree(writer->getPrecursorStruct(level)->dataD));
+ checkCudaErrors( cudaFree(writer->getPrecursorStruct(level)->bufferD));
+}
CudaMemoryManager::CudaMemoryManager(std::shared_ptr<Parameter> parameter) : parameter(parameter)
diff --git a/src/gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.h b/src/gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.h
index d410340d2de7797cf23a781a64d11f592d62a6fb..67347d005b1abb6ae6b6badf8f187a11b2bde15a 100644
--- a/src/gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.h
+++ b/src/gpu/VirtualFluids_GPU/GPU/CudaMemoryManager.h
@@ -18,8 +18,10 @@
class Parameter;
class PorousMedia;
-class ActuatorLine;
+class ActuatorFarm;
class Probe;
+class VelocitySetter;
+class PrecursorWriter;
class VIRTUALFLUIDS_GPU_EXPORT CudaMemoryManager
{
@@ -92,26 +94,20 @@ public:
//////////////////////////////////////////////////////////////////////////
//3D domain decomposition
virtual void cudaAllocProcessNeighborX(int lev, unsigned int processNeighbor);
- void cudaCopyProcessNeighborXFsHD(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsRecv,
- int streamIndex);
- void cudaCopyProcessNeighborXFsDH(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsSend,
- int streamIndex);
+ void cudaCopyProcessNeighborXFsHD(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsRecv);
+ void cudaCopyProcessNeighborXFsDH(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsSend);
virtual void cudaCopyProcessNeighborXIndex(int lev, unsigned int processNeighbor);
void cudaFreeProcessNeighborX(int lev, unsigned int processNeighbor);
//
virtual void cudaAllocProcessNeighborY(int lev, unsigned int processNeighbor);
- void cudaCopyProcessNeighborYFsHD(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsRecv,
- int streamIndex);
- void cudaCopyProcessNeighborYFsDH(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsSend,
- int streamIndex);
+ void cudaCopyProcessNeighborYFsHD(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsRecv);
+ void cudaCopyProcessNeighborYFsDH(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsSend);
virtual void cudaCopyProcessNeighborYIndex(int lev, unsigned int processNeighbor);
void cudaFreeProcessNeighborY(int lev, unsigned int processNeighbor);
//
virtual void cudaAllocProcessNeighborZ(int lev, unsigned int processNeighbor);
- void cudaCopyProcessNeighborZFsHD(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsRecv,
- int streamIndex);
- void cudaCopyProcessNeighborZFsDH(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsSend,
- int streamIndex);
+ void cudaCopyProcessNeighborZFsHD(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsRecv);
+ void cudaCopyProcessNeighborZFsDH(int lev, unsigned int processNeighbor, const unsigned int &memsizeFsSend);
virtual void cudaCopyProcessNeighborZIndex(int lev, unsigned int processNeighbor);
void cudaFreeProcessNeighborZ(int lev, unsigned int processNeighbor);
@@ -183,6 +179,13 @@ public:
void cudaCopyStressBC(int lev);
void cudaFreeStressBC(int lev);
+ void cudaAllocPrecursorBC(int lev);
+ void cudaAllocPrecursorData(int lev);
+ void cudaCopyPrecursorBC(int lev);
+ void cudaCopyPrecursorData(int lev);
+ void cudaFreePrecursorBC(int lev);
+ void cudaFreePrecursorData(int lev);
+
void cudaAllocWallModel(int lev, bool hasWallModelMonitor);
void cudaCopyWallModel(int lev, bool hasWallModelMonitor);
void cudaFreeWallModel(int lev, bool hasWallModelMonitor);
@@ -346,42 +349,44 @@ public:
void cudaCopyProcessNeighborADZIndex(int lev, unsigned int processNeighbor);
void cudaFreeProcessNeighborADZ(int lev, unsigned int processNeighbor);
- void cudaAllocFluidNodeIndices(int lev);
- void cudaCopyFluidNodeIndices(int lev);
- void cudaFreeFluidNodeIndices(int lev);
- void cudaAllocFluidNodeIndicesBorder(int lev);
- void cudaCopyFluidNodeIndicesBorder(int lev);
- void cudaFreeFluidNodeIndicesBorder(int lev);
-
- // Actuator Line
- void cudaAllocBladeRadii(ActuatorLine* actuatorLine);
- void cudaCopyBladeRadiiHtoD(ActuatorLine* actuatorLine);
- void cudaCopyBladeRadiiDtoH(ActuatorLine* actuatorLine);
- void cudaFreeBladeRadii(ActuatorLine* actuatorLine);
-
- void cudaAllocBladeCoords(ActuatorLine* actuatorLine);
- void cudaCopyBladeCoordsHtoD(ActuatorLine* actuatorLine);
- void cudaCopyBladeCoordsDtoH(ActuatorLine* actuatorLine);
- void cudaFreeBladeCoords(ActuatorLine* actuatorLine);
-
- void cudaAllocBladeIndices(ActuatorLine* actuatorLine);
- void cudaCopyBladeIndicesHtoD(ActuatorLine* actuatorLine);
- void cudaFreeBladeIndices(ActuatorLine* actuatorLine);
-
- void cudaAllocBladeVelocities(ActuatorLine* actuatorLine);
- void cudaCopyBladeVelocitiesHtoD(ActuatorLine* actuatorLine);
- void cudaCopyBladeVelocitiesDtoH(ActuatorLine* actuatorLine);
- void cudaFreeBladeVelocities(ActuatorLine* actuatorLine);
-
- void cudaAllocBladeForces(ActuatorLine* actuatorLine);
- void cudaCopyBladeForcesHtoD(ActuatorLine* actuatorLine);
- void cudaCopyBladeForcesDtoH(ActuatorLine* actuatorLine);
- void cudaFreeBladeForces(ActuatorLine* actuatorLine);
-
- void cudaAllocSphereIndices(ActuatorLine* actuatorLine);
- void cudaCopySphereIndicesHtoD(ActuatorLine* actuatorLine);
- void cudaFreeSphereIndices(ActuatorLine* actuatorLine);
-
+ void cudaAllocTaggedFluidNodeIndices(CollisionTemplate tag, int lev);
+ void cudaCopyTaggedFluidNodeIndices(CollisionTemplate tag, int lev);
+ void cudaFreeTaggedFluidNodeIndices(CollisionTemplate tag, int lev);
+
+ // ActuatorFarm
+ void cudaAllocBladeGeometries(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeGeometriesHtoD(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeGeometriesDtoH(ActuatorFarm* actuatorFarm);
+ void cudaFreeBladeGeometries(ActuatorFarm* actuatorFarm);
+
+ void cudaAllocBladeOrientations(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeOrientationsHtoD(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeOrientationsDtoH(ActuatorFarm* actuatorFarm);
+ void cudaFreeBladeOrientations(ActuatorFarm* actuatorFarm);
+
+ void cudaAllocBladeCoords(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeCoordsHtoD(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeCoordsDtoH(ActuatorFarm* actuatorFarm);
+ void cudaFreeBladeCoords(ActuatorFarm* actuatorFarm);
+
+ void cudaAllocBladeIndices(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeIndicesHtoD(ActuatorFarm* actuatorFarm);
+ void cudaFreeBladeIndices(ActuatorFarm* actuatorFarm);
+
+ void cudaAllocBladeVelocities(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeVelocitiesHtoD(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeVelocitiesDtoH(ActuatorFarm* actuatorFarm);
+ void cudaFreeBladeVelocities(ActuatorFarm* actuatorFarm);
+
+ void cudaAllocBladeForces(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeForcesHtoD(ActuatorFarm* actuatorFarm);
+ void cudaCopyBladeForcesDtoH(ActuatorFarm* actuatorFarm);
+ void cudaFreeBladeForces(ActuatorFarm* actuatorFarm);
+
+ void cudaAllocSphereIndices(ActuatorFarm* actuatorFarm);
+ void cudaCopySphereIndicesHtoD(ActuatorFarm* actuatorFarm);
+ void cudaFreeSphereIndices(ActuatorFarm* actuatorFarm);
+ // Probes
void cudaAllocProbeDistances(Probe* probe, int level);
void cudaCopyProbeDistancesHtoD(Probe* probe, int level);
void cudaCopyProbeDistancesDtoH(Probe* probe, int level);
@@ -402,6 +407,12 @@ public:
void cudaCopyProbeQuantitiesAndOffsetsDtoH(Probe* probe, int level);
void cudaFreeProbeQuantitiesAndOffsets(Probe* probe, int level);
+ //Precursor Writer
+ void cudaAllocPrecursorWriter(PrecursorWriter* writer, int level);
+ void cudaCopyPrecursorWriterIndicesHtoD(PrecursorWriter* writer, int level);
+ void cudaCopyPrecursorWriterOutputVariablesDtoH(PrecursorWriter* writer, int level);
+ void cudaFreePrecursorWriter(PrecursorWriter* writer, int level);
+
private:
std::shared_ptr<Parameter> parameter;
double memsizeGPU = 0.0;
diff --git a/src/gpu/VirtualFluids_GPU/GPU/GPU_Interface.h b/src/gpu/VirtualFluids_GPU/GPU/GPU_Interface.h
index ee987ae23402ef304220349db77084cc341ccd5a..ceb70fb123c52c282200137a00522ff2b9905f86 100644
--- a/src/gpu/VirtualFluids_GPU/GPU/GPU_Interface.h
+++ b/src/gpu/VirtualFluids_GPU/GPU/GPU_Interface.h
@@ -900,6 +900,8 @@ void QPressDevDirDepBot27(unsigned int numberOfThreads,
void QPressNoRhoDev27(LBMSimulationParameter* parameterDevice, QforBoundaryConditions* boundaryCondition);
+void QPressZeroRhoOutflowDev27(LBMSimulationParameter* parameterDevice, QforBoundaryConditions* boundaryCondition);
+
void QInflowScaleByPressDev27(LBMSimulationParameter* parameterDevice, QforBoundaryConditions* boundaryCondition);
void QPressDevOld27(unsigned int numberOfThreads,
@@ -1007,6 +1009,14 @@ void VelSchlaffer27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep);
+void QPrecursorDevCompZeroPress(LBMSimulationParameter* parameterDevice, QforPrecursorBoundaryConditions* boundaryCondition, real tRatio, real velocityRatio);
+
+void PrecursorDevEQ27(LBMSimulationParameter* parameterDevice, QforPrecursorBoundaryConditions* boundaryCondition, real tRatio, real velocityRatio);
+
+void PrecursorDevDistributions(LBMSimulationParameter* parameterDevice, QforPrecursorBoundaryConditions* boundaryCondition, real tRatio, real velocityRatio);
+
+void QPrecursorDevDistributions(LBMSimulationParameter* parameterDevice, QforPrecursorBoundaryConditions* boundaryCondition, real tRatio, real velocityRatio);
+
void QADDev7(unsigned int numberOfThreads,
real* DD,
real* DD7,
diff --git a/src/gpu/VirtualFluids_GPU/GPU/GPU_Kernels.cuh b/src/gpu/VirtualFluids_GPU/GPU/GPU_Kernels.cuh
index 94b9704b7ca57df4cd985f5aff9521b8a087b97f..a7b22c244da368b9591c5f528bb7bd5eaea91e50 100644
--- a/src/gpu/VirtualFluids_GPU/GPU/GPU_Kernels.cuh
+++ b/src/gpu/VirtualFluids_GPU/GPU/GPU_Kernels.cuh
@@ -1080,7 +1080,7 @@ __global__ void QPressDeviceDirDepBot27( real* rhoBC,
bool isEvenTimestep);
__global__ void QPressNoRhoDevice27( real* rhoBC,
- real* DD,
+ real* distributions,
int* k_Q,
int* k_N,
int numberOfBCnodes,
@@ -1088,8 +1088,23 @@ __global__ void QPressNoRhoDevice27( real* rhoBC,
unsigned int* neighborX,
unsigned int* neighborY,
unsigned int* neighborZ,
- unsigned int size_Mat,
- bool isEvenTimestep);
+ unsigned int numberOfLBnodes,
+ bool isEvenTimestep,
+ int direction);
+
+__global__ void QPressZeroRhoOutflowDevice27( real* rhoBC,
+ real* distributions,
+ int* k_Q,
+ int* k_N,
+ int numberOfBCnodes,
+ real om1,
+ unsigned int* neighborX,
+ unsigned int* neighborY,
+ unsigned int* neighborZ,
+ unsigned int numberOfLBnodes,
+ bool isEvenTimestep,
+ int direction,
+ real densityCorrectionFactor);
__global__ void QInflowScaleByPressDevice27( real* rhoBC,
real* DD,
@@ -1228,6 +1243,103 @@ __global__ void VelSchlaff27( int t,
unsigned int size_Mat,
bool isEvenTimestep);
+__global__ void QPrecursorDeviceCompZeroPress( int* subgridDistanceIndices,
+ int numberOfBCnodes,
+ int numberOfPrecursorNodes,
+ int sizeQ,
+ real omega,
+ real* distributions,
+ real* subgridDistances,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* neighborsNT,
+ uint* neighborsNB,
+ uint* neighborsST,
+ uint* neighborsSB,
+ real* weights0PP,
+ real* weights0PM,
+ real* weights0MP,
+ real* weights0MM,
+ real* vLast,
+ real* vCurrent,
+ real velocityX,
+ real velocityY,
+ real velocityZ,
+ real timeRatio,
+ real velocityRatio,
+ unsigned long long numberOfLBnodes,
+ bool isEvenTimestep);
+
+__global__ void PrecursorDeviceEQ27( int* subgridDistanceIndices,
+ int numberOfBCnodes,
+ int numberOfPrecursorNodes,
+ real omega,
+ real* distributions,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* neighborsNT,
+ uint* neighborsNB,
+ uint* neighborsST,
+ uint* neighborsSB,
+ real* weights0PP,
+ real* weights0PM,
+ real* weights0MP,
+ real* weights0MM,
+ real* vLast,
+ real* vCurrent,
+ real velocityX,
+ real velocityY,
+ real velocityZ,
+ real timeRatio,
+ real velocityRatio,
+ unsigned long long numberOfLBnodes,
+ bool isEvenTimestep);
+
+__global__ void PrecursorDeviceDistributions( int* subgridDistanceIndices,
+ int numberOfBCNodes,
+ int numberOfPrecursorNodes,
+ real* distributions,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* neighborsNT,
+ uint* neighborsNB,
+ uint* neighborsST,
+ uint* neighborsSB,
+ real* weights0PP,
+ real* weights0PM,
+ real* weights0MP,
+ real* weights0MM,
+ real* fsLast,
+ real* fsNext,
+ real timeRatio,
+ unsigned long long numberOfLBnodes,
+ bool isEvenTimestep);
+__global__ void QPrecursorDeviceDistributions( int* subgridDistanceIndices,
+ real* subgridDistances,
+ int sizeQ,
+ int numberOfBCNodes,
+ int numberOfPrecursorNodes,
+ real* distributions,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* neighborsNT,
+ uint* neighborsNB,
+ uint* neighborsST,
+ uint* neighborsSB,
+ real* weights0PP,
+ real* weights0PM,
+ real* weights0MP,
+ real* weights0MM,
+ real* fsLast,
+ real* fsNext,
+ real timeRatio,
+ unsigned long long numberOfLBnodes,
+ bool isEvenTimestep);
+
//Advection / Diffusion BCs
__global__ void QAD7( real* DD,
real* DD7,
diff --git a/src/gpu/VirtualFluids_GPU/GPU/LBMKernel.cu b/src/gpu/VirtualFluids_GPU/GPU/LBMKernel.cu
index 63fc5be0ebe5d4a26d4662ee8c0dddbc3098247a..363fb76626b256205346a290a42a1da40e1c7445 100644
--- a/src/gpu/VirtualFluids_GPU/GPU/LBMKernel.cu
+++ b/src/gpu/VirtualFluids_GPU/GPU/LBMKernel.cu
@@ -54,22 +54,9 @@ void KernelCasSP27( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Casc_SP_27<<< grid, threads >>>(s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_Kernel_Casc_SP_27<<< grid.grid, grid.threads >>>(s9,
bcMatD,
neighborX,
neighborY,
@@ -90,22 +77,9 @@ void KernelCasSPMS27( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Casc_SP_MS_27<<< grid, threads >>>(s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_Kernel_Casc_SP_MS_27<<< grid.grid, grid.threads >>>(s9,
bcMatD,
neighborX,
neighborY,
@@ -126,22 +100,9 @@ void KernelCasSPMSOHM27( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Casc_SP_MS_OHM_27<<< grid, threads >>>( s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_Kernel_Casc_SP_MS_OHM_27<<< grid.grid, grid.threads >>>( s9,
bcMatD,
neighborX,
neighborY,
@@ -165,22 +126,9 @@ void KernelKumCompSRTSP27(
real* forces,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Kum_New_Comp_SRT_SP_27 <<< grid, threads >>>(
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_Kernel_Kum_New_Comp_SRT_SP_27 <<< grid.grid, grid.threads >>>(
omega,
bcMatD,
neighborX,
@@ -209,22 +157,9 @@ void KernelKum1hSP27( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Kum_1h_SP_27<<< grid, threads >>>(omega,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_Kernel_Kum_1h_SP_27<<< grid.grid, grid.threads >>>(omega,
deltaPhi,
angularVelocity,
bcMatD,
@@ -250,22 +185,9 @@ void KernelCascadeSP27( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Cascade_SP_27<<< grid, threads >>>(s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_Kernel_Cascade_SP_27<<< grid.grid, grid.threads >>>(s9,
bcMatD,
neighborX,
neighborY,
@@ -286,22 +208,10 @@ void KernelKumNewSP27( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Kum_New_SP_27<<< grid, threads >>>(s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ LB_Kernel_Kum_New_SP_27<<< grid.grid, grid.threads >>>(s9,
bcMatD,
neighborX,
neighborY,
@@ -329,22 +239,9 @@ void KernelKumNewCompSP27(unsigned int numberOfThreads,
//dim3 grid(Grid, 1, 1);
//dim3 threads(numberOfThreads, 1, 1 );
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2, 1);
- dim3 threads(numberOfThreads, 1, 1);
-
- //LB_Kernel_Kum_New_Comp_SP_27<<< grid, threads >>>( s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ //LB_Kernel_Kum_New_Comp_SP_27<<< grid.grid, grid.threads >>>( s9,
// bcMatD,
// neighborX,
// neighborY,
@@ -375,22 +272,10 @@ void CumulantOnePreconditionedErrorDiffusionChimCompSP27(unsigned int numberOfTh
//dim3 grid(Grid, 1, 1);
//dim3 threads(numberOfThreads, 1, 1 );
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2, 1);
- dim3 threads(numberOfThreads, 1, 1);
-
- Cumulant_One_preconditioned_errorDiffusion_chim_Comp_SP_27 <<< grid, threads >>>( s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ Cumulant_One_preconditioned_errorDiffusion_chim_Comp_SP_27 <<< grid.grid, grid.threads >>>( s9,
bcMatD,
neighborX,
neighborY,
@@ -420,22 +305,10 @@ void CumulantOnePreconditionedChimCompSP27( unsigned int numberOfThreads,
//dim3 grid(Grid, 1, 1);
//dim3 threads(numberOfThreads, 1, 1 );
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2, 1);
- dim3 threads(numberOfThreads, 1, 1);
-
- Cumulant_One_preconditioned_chim_Comp_SP_27 <<< grid, threads >>>( s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ Cumulant_One_preconditioned_chim_Comp_SP_27 <<< grid.grid, grid.threads >>>( s9,
bcMatD,
neighborX,
neighborY,
@@ -465,22 +338,10 @@ void CumulantOneChimCompSP27(unsigned int numberOfThreads,
//dim3 grid(Grid, 1, 1);
//dim3 threads(numberOfThreads, 1, 1 );
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2, 1);
- dim3 threads(numberOfThreads, 1, 1);
-
- Cumulant_One_chim_Comp_SP_27 <<< grid, threads >>>( s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ Cumulant_One_chim_Comp_SP_27 <<< grid.grid, grid.threads >>>( s9,
bcMatD,
neighborX,
neighborY,
@@ -506,22 +367,10 @@ void KernelKumIsoTestSP27(unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Kum_IsoTest_SP_27<<< grid, threads >>>(s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ LB_Kernel_Kum_IsoTest_SP_27<<< grid.grid, grid.threads >>>(s9,
bcMatD,
neighborX,
neighborY,
@@ -545,22 +394,10 @@ void KernelKumCompSP27( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_Kum_Comp_SP_27<<< grid, threads >>>(s9,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ LB_Kernel_Kum_Comp_SP_27<<< grid.grid, grid.threads >>>(s9,
bcMatD,
neighborX,
neighborY,
@@ -587,22 +424,10 @@ void KernelPMCumOneCompSP27(unsigned int numberOfThreads,
unsigned int* nodeIdsPorousMedia,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2, 1);
- dim3 threads(numberOfThreads, 1, 1);
-
- LB_Kernel_PM_Cum_One_Comp_SP_27 <<< grid, threads >>>(omega,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ LB_Kernel_PM_Cum_One_Comp_SP_27 <<< grid.grid, grid.threads >>>(omega,
neighborX,
neighborY,
neighborZ,
@@ -642,22 +467,10 @@ void KernelWaleBySoniMalavCumAA2016CompSP27(
//dim3 grid(Grid, 1, 1);
//dim3 threads(numberOfThreads, 1, 1 );
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2, 1);
- dim3 threads(numberOfThreads, 1, 1);
-
- LB_Kernel_WaleBySoniMalav_Cum_AA2016_Comp_SP_27 << < grid, threads >> >(
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ LB_Kernel_WaleBySoniMalav_Cum_AA2016_Comp_SP_27 << < grid.grid, grid.threads >> >(
s9,
bcMatD,
neighborX,
@@ -687,22 +500,9 @@ void KernelADincomp7( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_AD_Incomp_7<<< grid, threads >>>( diffusivity,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_Kernel_AD_Incomp_7<<< grid.grid, grid.threads >>>( diffusivity,
bcMatD,
neighborX,
neighborY,
@@ -725,22 +525,9 @@ void KernelADincomp27( unsigned int numberOfThreads,
int size_Mat,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LB_Kernel_AD_Incomp_27<<< grid, threads >>>( diffusivity,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_Kernel_AD_Incomp_27<<< grid.grid, grid.threads >>>( diffusivity,
bcMatD,
neighborX,
neighborY,
@@ -749,7 +536,7 @@ void KernelADincomp27( unsigned int numberOfThreads,
DD27,
size_Mat,
EvenOrOdd);
- getLastCudaError("LB_Kernel_AD_Incomp_27 execution failed");
+ getLastCudaError("LB_Kernel_AD_Incomp_27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void Init27( int myid,
@@ -771,7 +558,7 @@ void Init27( int myid,
dim3 threads ( grid_nx, 1, 1 );
dim3 grid ( grid_ny, grid_nz ); // Gitter fuer Kollision und Propagation
- LBInit27<<< grid, threads >>> ( myid,
+ LBInit27<<< grid, threads >>> ( myid,
numprocs,
u0,
geoD,
@@ -786,7 +573,7 @@ void Init27( int myid,
DD,
level,
maxlevel);
- getLastCudaError("LBInit27 execution failed");
+ getLastCudaError("LBInit27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void InitNonEqPartSP27( unsigned int numberOfThreads,
@@ -804,22 +591,9 @@ void InitNonEqPartSP27( unsigned int numberOfThreads,
real omega,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBInitNonEqPartSP27<<< grid, threads >>>( neighborX,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBInitNonEqPartSP27<<< grid.grid, grid.threads >>>( neighborX,
neighborY,
neighborZ,
neighborWSB,
@@ -832,7 +606,7 @@ void InitNonEqPartSP27( unsigned int numberOfThreads,
DD,
omega,
EvenOrOdd);
- getLastCudaError("LBInitNonEqPartSP27 execution failed");
+ getLastCudaError("LBInitNonEqPartSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void InitThS7( unsigned int numberOfThreads,
@@ -848,22 +622,9 @@ void InitThS7( unsigned int numberOfThreads,
real* DD7,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- InitAD7<<< grid, threads >>>( neighborX,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ InitAD7<<< grid.grid, grid.threads >>>( neighborX,
neighborY,
neighborZ,
geoD,
@@ -874,7 +635,7 @@ void InitThS7( unsigned int numberOfThreads,
size_Mat,
DD7,
EvenOrOdd);
- getLastCudaError("InitAD7 execution failed");
+ getLastCudaError("InitAD7 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void InitADDev27( unsigned int numberOfThreads,
@@ -890,22 +651,9 @@ void InitADDev27( unsigned int numberOfThreads,
real* DD27,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- InitAD27<<< grid, threads >>>(neighborX,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ InitAD27<<< grid.grid, grid.threads >>>(neighborX,
neighborY,
neighborZ,
geoD,
@@ -916,7 +664,7 @@ void InitADDev27( unsigned int numberOfThreads,
size_Mat,
DD27,
EvenOrOdd);
- getLastCudaError("InitAD27 execution failed");
+ getLastCudaError("InitAD27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void PostProcessorF3_2018Fehlberg(
@@ -937,22 +685,9 @@ void PostProcessorF3_2018Fehlberg(
real* forces,
bool EvenOrOdd)
{
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1);
-
- LB_PostProcessor_F3_2018_Fehlberg <<< grid, threads >>> ( omega,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LB_PostProcessor_F3_2018_Fehlberg <<< grid.grid, grid.threads >>> ( omega,
bcMatD,
neighborX,
neighborY,
@@ -967,7 +702,7 @@ void PostProcessorF3_2018Fehlberg(
level,
forces,
EvenOrOdd);
- getLastCudaError("LB_PostProcessor_F3_2018_Fehlberg execution failed");
+ getLastCudaError("LB_PostProcessor_F3_2018_Fehlberg execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcMac27( real* vxD,
@@ -988,7 +723,7 @@ void CalcMac27( real* vxD,
dim3 threads ( grid_nx, 1, 1 );
dim3 grid ( grid_ny, grid_nz );
- LBCalcMac27<<< grid, threads >>> ( vxD,
+ LBCalcMac27<<< grid, threads >>> ( vxD,
vyD,
vzD,
rhoD,
@@ -999,7 +734,7 @@ void CalcMac27( real* vxD,
size_Mat,
DD,
isEvenTimestep);
- getLastCudaError("LBCalcMac27 execution failed");
+ getLastCudaError("LBCalcMac27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcMacSP27( real* vxD,
@@ -1016,22 +751,9 @@ void CalcMacSP27( real* vxD,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalcMacSP27<<< grid, threads >>> ( vxD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalcMacSP27<<< grid.grid, grid.threads >>> ( vxD,
vyD,
vzD,
rhoD,
@@ -1043,7 +765,7 @@ void CalcMacSP27( real* vxD,
size_Mat,
DD,
isEvenTimestep);
- getLastCudaError("LBCalcMacSP27 execution failed");
+ getLastCudaError("LBCalcMacSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcMacCompSP27( real* vxD,
@@ -1060,22 +782,9 @@ void CalcMacCompSP27( real* vxD,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalcMacCompSP27<<< grid, threads >>> ( vxD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalcMacCompSP27<<< grid.grid, grid.threads >>> ( vxD,
vyD,
vzD,
rhoD,
@@ -1087,7 +796,7 @@ void CalcMacCompSP27( real* vxD,
size_Mat,
DD,
isEvenTimestep);
- getLastCudaError("LBCalcMacSP27 execution failed");
+ getLastCudaError("LBCalcMacSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcMacThS7( real* Conc,
@@ -1100,22 +809,9 @@ void CalcMacThS7( real* Conc,
real* DD7,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- CalcConc7<<< grid, threads >>> (Conc,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ CalcConc7<<< grid.grid, grid.threads >>> (Conc,
geoD,
neighborX,
neighborY,
@@ -1123,7 +819,7 @@ void CalcMacThS7( real* Conc,
size_Mat,
DD7,
isEvenTimestep);
- getLastCudaError("CalcConc7 execution failed");
+ getLastCudaError("CalcConc7 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void PlaneConcThS7(real* Conc,
@@ -1138,22 +834,9 @@ void PlaneConcThS7(real* Conc,
real* DD7,
bool isEvenTimestep)
{
- int Grid = (numberOfPointskPC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- GetPlaneConc7<<< grid, threads >>> ( Conc,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfPointskPC);
+
+ GetPlaneConc7<<< grid.grid, grid.threads >>> ( Conc,
kPC,
numberOfPointskPC,
geoD,
@@ -1163,7 +846,7 @@ void PlaneConcThS7(real* Conc,
size_Mat,
DD7,
isEvenTimestep);
- getLastCudaError("GetPlaneConc7 execution failed");
+ getLastCudaError("GetPlaneConc7 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void PlaneConcThS27(real* Conc,
@@ -1178,22 +861,9 @@ void PlaneConcThS27(real* Conc,
real* DD27,
bool isEvenTimestep)
{
- int Grid = (numberOfPointskPC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- GetPlaneConc27<<< grid, threads >>> ( Conc,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfPointskPC);
+
+ GetPlaneConc27<<< grid.grid, grid.threads >>> ( Conc,
kPC,
numberOfPointskPC,
geoD,
@@ -1203,7 +873,7 @@ void PlaneConcThS27(real* Conc,
size_Mat,
DD27,
isEvenTimestep);
- getLastCudaError("GetPlaneConc27 execution failed");
+ getLastCudaError("GetPlaneConc27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcConcentration27( unsigned int numberOfThreads,
@@ -1216,22 +886,9 @@ void CalcConcentration27( unsigned int numberOfThreads,
real* DD27,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- CalcConc27<<< grid, threads >>> ( Conc,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ CalcConc27<<< grid.grid, grid.threads >>> ( Conc,
geoD,
neighborX,
neighborY,
@@ -1239,7 +896,7 @@ void CalcConcentration27( unsigned int numberOfThreads,
size_Mat,
DD27,
isEvenTimestep);
- getLastCudaError("CalcConc27 execution failed");
+ getLastCudaError("CalcConc27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcMedSP27( real* vxD,
@@ -1256,22 +913,9 @@ void CalcMedSP27( real* vxD,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalcMedSP27<<< grid, threads >>> ( vxD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalcMedSP27<<< grid.grid, grid.threads >>> ( vxD,
vyD,
vzD,
rhoD,
@@ -1283,7 +927,7 @@ void CalcMedSP27( real* vxD,
size_Mat,
DD,
isEvenTimestep);
- getLastCudaError("LBCalcMedSP27 execution failed");
+ getLastCudaError("LBCalcMedSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcMedCompSP27( real* vxD,
@@ -1300,22 +944,9 @@ void CalcMedCompSP27( real* vxD,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalcMedCompSP27<<< grid, threads >>> ( vxD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalcMedCompSP27<<< grid.grid, grid.threads >>> ( vxD,
vyD,
vzD,
rhoD,
@@ -1327,7 +958,7 @@ void CalcMedCompSP27( real* vxD,
size_Mat,
DD,
isEvenTimestep);
- getLastCudaError("LBCalcMedSP27 execution failed");
+ getLastCudaError("LBCalcMedSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcMedCompAD27(
@@ -1347,22 +978,9 @@ void CalcMedCompAD27(
real* DD_AD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1);
-
- LBCalcMedCompAD27 <<< grid, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalcMedCompAD27 <<< grid.grid, grid.threads >>> (
vxD,
vyD,
vzD,
@@ -1394,22 +1012,9 @@ void CalcMacMedSP27( real* vxD,
unsigned int numberOfThreads,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalcMacMedSP27<<< grid, threads >>> ( vxD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalcMacMedSP27<<< grid.grid, grid.threads >>> ( vxD,
vyD,
vzD,
rhoD,
@@ -1421,7 +1026,7 @@ void CalcMacMedSP27( real* vxD,
tdiff,
size_Mat,
isEvenTimestep);
- getLastCudaError("LBCalcMacMedSP27 execution failed");
+ getLastCudaError("LBCalcMacMedSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void ResetMedianValuesSP27(
@@ -1434,22 +1039,10 @@ void ResetMedianValuesSP27(
unsigned int numberOfThreads,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1);
-
- LBResetMedianValuesSP27 << < grid, threads >> > (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+
+ LBResetMedianValuesSP27 << < grid.grid, grid.threads >> > (
vxD,
vyD,
vzD,
@@ -1471,22 +1064,9 @@ void ResetMedianValuesAD27(
unsigned int numberOfThreads,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1);
-
- LBResetMedianValuesAD27 << < grid, threads >> > (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBResetMedianValuesAD27 << < grid.grid, grid.threads >> > (
vxD,
vyD,
vzD,
@@ -1512,22 +1092,9 @@ void Calc2ndMomentsIncompSP27(real* kxyFromfcNEQ,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalc2ndMomentsIncompSP27<<< grid, threads >>> ( kxyFromfcNEQ,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalc2ndMomentsIncompSP27<<< grid.grid, grid.threads >>> ( kxyFromfcNEQ,
kyzFromfcNEQ,
kxzFromfcNEQ,
kxxMyyFromfcNEQ,
@@ -1539,7 +1106,7 @@ void Calc2ndMomentsIncompSP27(real* kxyFromfcNEQ,
size_Mat,
DD,
isEvenTimestep);
- getLastCudaError("LBCalc2ndMomentsIncompSP27 execution failed");
+ getLastCudaError("LBCalc2ndMomentsIncompSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void Calc2ndMomentsCompSP27( real* kxyFromfcNEQ,
@@ -1556,22 +1123,9 @@ void Calc2ndMomentsCompSP27( real* kxyFromfcNEQ,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalc2ndMomentsCompSP27<<< grid, threads >>> (kxyFromfcNEQ,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalc2ndMomentsCompSP27<<< grid.grid, grid.threads >>> (kxyFromfcNEQ,
kyzFromfcNEQ,
kxzFromfcNEQ,
kxxMyyFromfcNEQ,
@@ -1583,7 +1137,7 @@ void Calc2ndMomentsCompSP27( real* kxyFromfcNEQ,
size_Mat,
DD,
isEvenTimestep);
- getLastCudaError("LBCalc2ndMomentsCompSP27 execution failed");
+ getLastCudaError("LBCalc2ndMomentsCompSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void Calc3rdMomentsIncompSP27(real* CUMbbb,
@@ -1602,22 +1156,9 @@ void Calc3rdMomentsIncompSP27(real* CUMbbb,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalc3rdMomentsIncompSP27<<< grid, threads >>> ( CUMbbb,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalc3rdMomentsIncompSP27<<< grid.grid, grid.threads >>> ( CUMbbb,
CUMabc,
CUMbac,
CUMbca,
@@ -1631,7 +1172,7 @@ void Calc3rdMomentsIncompSP27(real* CUMbbb,
DD,
size_Mat,
isEvenTimestep);
- getLastCudaError("LBCalc3rdMomentsIncompSP27 execution failed");
+ getLastCudaError("LBCalc3rdMomentsIncompSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void Calc3rdMomentsCompSP27( real* CUMbbb,
@@ -1650,22 +1191,9 @@ void Calc3rdMomentsCompSP27( real* CUMbbb,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalc3rdMomentsCompSP27<<< grid, threads >>> (CUMbbb,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalc3rdMomentsCompSP27<<< grid.grid, grid.threads >>> (CUMbbb,
CUMabc,
CUMbac,
CUMbca,
@@ -1679,7 +1207,7 @@ void Calc3rdMomentsCompSP27( real* CUMbbb,
DD,
size_Mat,
isEvenTimestep);
- getLastCudaError("LBCalc3rdMomentsCompSP27 execution failed");
+ getLastCudaError("LBCalc3rdMomentsCompSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcHigherMomentsIncompSP27(real* CUMcbb,
@@ -1701,22 +1229,9 @@ void CalcHigherMomentsIncompSP27(real* CUMcbb,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalcHigherMomentsIncompSP27<<< grid, threads >>> (CUMcbb,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalcHigherMomentsIncompSP27<<< grid.grid, grid.threads >>> (CUMcbb,
CUMbcb,
CUMbbc,
CUMcca,
@@ -1733,7 +1248,7 @@ void CalcHigherMomentsIncompSP27(real* CUMcbb,
DD,
size_Mat,
isEvenTimestep);
- getLastCudaError("LBCalcHigherMomentsIncompSP27 execution failed");
+ getLastCudaError("LBCalcHigherMomentsIncompSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void CalcHigherMomentsCompSP27( real* CUMcbb,
@@ -1755,22 +1270,9 @@ void CalcHigherMomentsCompSP27( real* CUMcbb,
real* DD,
bool isEvenTimestep)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalcHigherMomentsCompSP27<<< grid, threads >>> ( CUMcbb,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+
+ LBCalcHigherMomentsCompSP27<<< grid.grid, grid.threads >>> ( CUMcbb,
CUMbcb,
CUMbbc,
CUMcca,
@@ -1787,7 +1289,7 @@ void CalcHigherMomentsCompSP27( real* CUMcbb,
DD,
size_Mat,
isEvenTimestep);
- getLastCudaError("LBCalcHigherMomentsCompSP27 execution failed");
+ getLastCudaError("LBCalcHigherMomentsCompSP27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void LBCalcMeasurePoints27(real* vxMP,
@@ -1807,22 +1309,9 @@ void LBCalcMeasurePoints27(real* vxMP,
unsigned int numberOfThreads,
bool isEvenTimestep)
{
- int Grid = (numberOfPointskMP / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBCalcMeasurePoints<<< grid, threads >>> (vxMP,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfPointskMP);
+
+ LBCalcMeasurePoints<<< grid.grid, grid.threads >>> (vxMP,
vyMP,
vzMP,
rhoMP,
@@ -1837,7 +1326,7 @@ void LBCalcMeasurePoints27(real* vxMP,
size_Mat,
DD,
isEvenTimestep);
- getLastCudaError("LBCalcMeasurePoints execution failed");
+ getLastCudaError("LBCalcMeasurePoints execution failed");
}
//////////////////////////////////////////////////////////////////////////
void BcPress27( int nx,
@@ -1853,10 +1342,10 @@ void BcPress27( int nx,
unsigned int size_Mat,
bool isEvenTimestep)
{
- dim3 threads ( grid_nx, 1, 1 );
- dim3 grid ( grid_ny, 1 );
+ dim3 threads ( grid_nx, 1, 1 );
+ dim3 grid ( grid_ny, 1 );
- LB_BC_Press_East27<<< grid, threads >>> ( nx,
+ LB_BC_Press_East27<<< grid, threads >>> ( nx,
ny,
tz,
bcMatD,
@@ -1866,7 +1355,7 @@ void BcPress27( int nx,
DD,
size_Mat,
isEvenTimestep);
- getLastCudaError("LB_BC_Press_East27 execution failed");
+ getLastCudaError("LB_BC_Press_East27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void BcVel27(int nx,
@@ -1885,10 +1374,10 @@ void BcVel27(int nx,
real u0x,
real om)
{
- dim3 threads ( grid_nx, 1, 1 );
- dim3 grid ( grid_ny, 1 );
+ dim3 threads ( grid_nx, 1, 1 );
+ dim3 grid ( grid_ny, 1 );
- LB_BC_Vel_West_27<<< grid, threads >>> ( nx,
+ LB_BC_Vel_West_27<<< grid, threads >>> ( nx,
ny,
nz,
itz,
@@ -1903,7 +1392,7 @@ void BcVel27(int nx,
grid_nx,
grid_ny,
om);
- getLastCudaError("LB_BC_Vel_West_27 execution failed");
+ getLastCudaError("LB_BC_Vel_West_27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QADPressDev7( unsigned int numberOfThreads,
@@ -1922,22 +1411,9 @@ void QADPressDev7( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADPress7<<< gridQ, threads >>>( DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADPress7<<< grid.grid, grid.threads >>>( DD,
DD7,
temp,
velo,
@@ -1951,7 +1427,7 @@ void QADPressDev7( unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QADPress7 execution failed");
+ getLastCudaError("QADPress7 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QADPressDev27(unsigned int numberOfThreads,
@@ -1970,22 +1446,9 @@ void QADPressDev27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADPress27<<< gridQ, threads >>>( DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADPress27<<< grid.grid, grid.threads >>>( DD,
DD27,
temp,
velo,
@@ -1999,7 +1462,7 @@ void QADPressDev27(unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QADPress27 execution failed");
+ getLastCudaError("QADPress27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QADPressNEQNeighborDev27(
@@ -2017,22 +1480,9 @@ void QADPressNEQNeighborDev27(
)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADPressNEQNeighbor27<<< gridQ, threads >>>(
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADPressNEQNeighbor27<<< grid.grid, grid.threads >>>(
DD,
DD27,
k_Q,
@@ -2044,7 +1494,7 @@ void QADPressNEQNeighborDev27(
size_Mat,
isEvenTimestep
);
- getLastCudaError("QADPressNEQNeighbor27 execution failed");
+ getLastCudaError("QADPressNEQNeighbor27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QADVelDev7(unsigned int numberOfThreads,
@@ -2063,22 +1513,9 @@ void QADVelDev7(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADVel7<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADVel7<<< grid.grid, grid.threads >>> (
DD,
DD7,
temp,
@@ -2093,7 +1530,7 @@ void QADVelDev7(unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QADVel7 execution failed");
+ getLastCudaError("QADVel7 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QADVelDev27( unsigned int numberOfThreads,
@@ -2112,22 +1549,9 @@ void QADVelDev27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADVel27<<< gridQ, threads >>> ( DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADVel27<<< grid.grid, grid.threads >>> ( DD,
DD27,
temp,
velo,
@@ -2159,22 +1583,9 @@ void QADDev7(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QAD7<<< gridQ, threads >>> ( DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QAD7<<< grid.grid, grid.threads >>> ( DD,
DD7,
temp,
diffusivity,
@@ -2242,11 +1653,9 @@ void ADSlipVelDevComp(
uint size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads) + 1;
- dim3 gridQ(Grid, 1, 1);
- dim3 threads(numberOfThreads, 1, 1);
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
- AD_SlipVelDeviceComp << < gridQ, threads >> > (
+ AD_SlipVelDeviceComp << < grid.grid, grid.threads >> > (
normalX,
normalY,
normalZ,
@@ -2280,22 +1689,9 @@ void QADDirichletDev27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADDirichlet27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADDirichlet27<<< grid.grid, grid.threads >>> (
DD,
DD27,
temp,
@@ -2327,22 +1723,9 @@ void QADBBDev27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADBB27<<< gridQ, threads >>> ( DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADBB27<<< grid.grid, grid.threads >>> ( DD,
DD27,
temp,
diffusivity,
@@ -2373,22 +1756,9 @@ void QNoSlipADincompDev7(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QNoSlipADincomp7<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QNoSlipADincomp7<<< grid.grid, grid.threads >>> (
DD,
DD7,
temp,
@@ -2420,22 +1790,9 @@ void QNoSlipADincompDev27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QNoSlipADincomp27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QNoSlipADincomp27<<< grid.grid, grid.threads >>> (
DD,
DD27,
temp,
@@ -2468,24 +1825,10 @@ void QADVeloIncompDev7( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADVeloIncomp7<<< gridQ, threads >>> (
- DD,
- DD7,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADVeloIncomp7<<< grid.grid, grid.threads >>> ( DD,
+ DD7,
temp,
velo,
diffusivity,
@@ -2517,22 +1860,9 @@ void QADVeloIncompDev27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADVeloIncomp27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADVeloIncomp27<<< grid.grid, grid.threads >>> (
DD,
DD27,
temp,
@@ -2566,22 +1896,9 @@ void QADPressIncompDev7( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADPressIncomp7<<< gridQ, threads >>>(
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADPressIncomp7<<< grid.grid, grid.threads >>>(
DD,
DD7,
temp,
@@ -2615,24 +1932,10 @@ void QADPressIncompDev27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QADPressIncomp27<<< gridQ, threads >>>(
- DD,
- DD27,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QADPressIncomp27<<< grid.grid, grid.threads >>>(DD,
+ DD27,
temp,
velo,
diffusivity,
@@ -2701,22 +2004,9 @@ void QDevCompThinWalls27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QDeviceCompThinWallsPartOne27 <<< gridQ, threads >>> (DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QDeviceCompThinWallsPartOne27 <<< grid.grid, grid.threads >>> (DD,
k_Q,
QQ,
numberOfBCnodes,
@@ -2728,7 +2018,7 @@ void QDevCompThinWalls27(unsigned int numberOfThreads,
isEvenTimestep);
getLastCudaError("QDeviceCompThinWallsPartOne27 execution failed");
- QThinWallsPartTwo27 <<< gridQ, threads >>> ( DD,
+ QThinWallsPartTwo27 <<< grid.grid, grid.threads >>> ( DD,
k_Q,
QQ,
numberOfBCnodes,
@@ -2774,22 +2064,9 @@ void QDevIncompHighNu27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QDeviceIncompHighNu27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QDeviceIncompHighNu27<<< grid.grid, grid.threads >>> (
DD,
k_Q,
QQ,
@@ -2815,22 +2092,9 @@ void QDevCompHighNu27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QDeviceCompHighNu27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QDeviceCompHighNu27<<< grid.grid, grid.threads >>> (
DD,
k_Q,
QQ,
@@ -2880,22 +2144,9 @@ void QVelDeviceCouette27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QVelDevCouette27<<< gridQ, threads >>> ( vx,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QVelDevCouette27<<< grid.grid, grid.threads >>> ( vx,
vy,
vz,
DD,
@@ -2933,22 +2184,9 @@ void QVelDevice1h27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QVelDev1h27<<< gridQ, threads >>> (nx,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QVelDev1h27<<< grid.grid, grid.threads >>> (nx,
ny,
vx,
vy,
@@ -3010,22 +2248,9 @@ void QVelDevCompPlusSlip27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QVelDeviceCompPlusSlip27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QVelDeviceCompPlusSlip27<<< grid.grid, grid.threads >>> (
vx,
vy,
vz,
@@ -3081,22 +2306,9 @@ void QVelDevCompThinWalls27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QVelDeviceCompThinWallsPartOne27<<< gridQ, threads >>> (vx,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QVelDeviceCompThinWallsPartOne27<<< grid.grid, grid.threads >>> (vx,
vy,
vz,
DD,
@@ -3111,7 +2323,7 @@ void QVelDevCompThinWalls27(unsigned int numberOfThreads,
isEvenTimestep);
getLastCudaError("QVelDeviceCompThinWallsPartOne27 execution failed");
- QThinWallsPartTwo27 <<< gridQ, threads >>> (
+ QThinWallsPartTwo27 <<< grid.grid, grid.threads >>> (
DD,
k_Q,
QQ,
@@ -3163,22 +2375,9 @@ void QVelDevIncompHighNu27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QVelDeviceIncompHighNu27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QVelDeviceIncompHighNu27<<< grid.grid, grid.threads >>> (
vx,
vy,
vz,
@@ -3210,22 +2409,9 @@ void QVelDevCompHighNu27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QVelDeviceCompHighNu27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QVelDeviceCompHighNu27<<< grid.grid, grid.threads >>> (
vx,
vy,
vz,
@@ -3256,22 +2442,9 @@ void QVeloDevEQ27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QVeloDeviceEQ27<<< gridQ, threads >>> (VeloX,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QVeloDeviceEQ27<<< grid.grid, grid.threads >>> (VeloX,
VeloY,
VeloZ,
DD,
@@ -3301,22 +2474,9 @@ void QVeloStreetDevEQ27(
uint size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfStreetNodes / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1);
-
- QVeloStreetDeviceEQ27 << < gridQ, threads >> > (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfStreetNodes);
+
+ QVeloStreetDeviceEQ27 << < grid.grid, grid.threads >> > (
veloXfraction,
veloYfraction,
naschVelo,
@@ -3355,7 +2515,7 @@ void QSlipDevCompTurbulentViscosity27(LBMSimulationParameter* parameterDevice, Q
{
dim3 grid = vf::cuda::getCudaGrid( parameterDevice->numberofthreads, boundaryCondition->numberOfBCnodes);
dim3 threads(parameterDevice->numberofthreads, 1, 1 );
-
+
QSlipDeviceComp27TurbViscosity<<< grid, threads >>> (
parameterDevice->distributions.f[0],
boundaryCondition->k,
@@ -3395,7 +2555,7 @@ void QSlipDevComp27(LBMSimulationParameter* parameterDevice, QforBoundaryConditi
{
dim3 grid = vf::cuda::getCudaGrid( parameterDevice->numberofthreads, boundaryCondition->numberOfBCnodes);
dim3 threads(parameterDevice->numberofthreads, 1, 1 );
-
+
QSlipDeviceComp27<<< grid, threads >>> (
parameterDevice->distributions.f[0],
boundaryCondition->k,
@@ -3444,22 +2604,9 @@ void QSlipGeomDevComp27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QSlipGeomDeviceComp27<<< gridQ, threads >>> (DD,
+ vf::cuda::CudaGrid grid(numberOfThreads, numberOfBCnodes);
+
+ QSlipGeomDeviceComp27<<< grid.grid, grid.threads >>> (DD,
k_Q,
QQ,
numberOfBCnodes,
@@ -3472,7 +2619,7 @@ void QSlipGeomDevComp27(unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QSlipGeomDeviceComp27 execution failed");
+ getLastCudaError("QSlipGeomDeviceComp27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QSlipNormDevComp27(unsigned int numberOfThreads,
@@ -3490,22 +2637,9 @@ void QSlipNormDevComp27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QSlipNormDeviceComp27<<< gridQ, threads >>> (DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QSlipNormDeviceComp27<<< grid.grid, grid.threads >>> (DD,
k_Q,
QQ,
numberOfBCnodes,
@@ -3676,22 +2810,9 @@ void QPressDevAntiBB27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QPressDeviceAntiBB27<<< gridQ, threads >>>( rhoBC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QPressDeviceAntiBB27<<< grid.grid, grid.threads >>>( rhoBC,
vx,
vy,
vz,
@@ -3705,7 +2826,7 @@ void QPressDevAntiBB27( unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QPressDeviceAntiBB27 execution failed");
+ getLastCudaError("QPressDeviceAntiBB27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QPressDevFixBackflow27( unsigned int numberOfThreads,
@@ -3720,22 +2841,9 @@ void QPressDevFixBackflow27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QPressDeviceFixBackflow27<<< gridQ, threads >>> ( rhoBC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QPressDeviceFixBackflow27<<< grid.grid, grid.threads >>> ( rhoBC,
DD,
k_Q,
numberOfBCnodes,
@@ -3745,7 +2853,7 @@ void QPressDevFixBackflow27( unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QPressDeviceFixBackflow27 execution failed");
+ getLastCudaError("QPressDeviceFixBackflow27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QPressDevDirDepBot27( unsigned int numberOfThreads,
@@ -3760,22 +2868,9 @@ void QPressDevDirDepBot27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QPressDeviceDirDepBot27<<< gridQ, threads >>> ( rhoBC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QPressDeviceDirDepBot27<<< grid.grid, grid.threads >>> ( rhoBC,
DD,
k_Q,
numberOfBCnodes,
@@ -3785,7 +2880,7 @@ void QPressDevDirDepBot27( unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QPressDeviceDirDepBot27 execution failed");
+ getLastCudaError("QPressDeviceDirDepBot27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QPressNoRhoDev27(LBMSimulationParameter* parameterDevice, QforBoundaryConditions* boundaryCondition)
@@ -3804,10 +2899,33 @@ void QPressNoRhoDev27(LBMSimulationParameter* parameterDevice, QforBoundaryCondi
parameterDevice->neighborY,
parameterDevice->neighborZ,
parameterDevice->numberOfNodes,
- parameterDevice->isEvenTimestep);
+ parameterDevice->isEvenTimestep,
+ vf::lbm::dir::DIR_P00);
getLastCudaError("QPressNoRhoDevice27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
+void QPressZeroRhoOutflowDev27(LBMSimulationParameter* parameterDevice, QforBoundaryConditions* boundaryCondition)
+{
+ dim3 grid = vf::cuda::getCudaGrid( parameterDevice->numberofthreads, boundaryCondition->numberOfBCnodes);
+ dim3 threads(parameterDevice->numberofthreads, 1, 1 );
+
+ QPressZeroRhoOutflowDevice27<<< grid, threads >>> (
+ boundaryCondition->RhoBC,
+ parameterDevice->distributions.f[0],
+ boundaryCondition->k,
+ boundaryCondition->kN,
+ boundaryCondition->numberOfBCnodes,
+ parameterDevice->omega,
+ parameterDevice->neighborX,
+ parameterDevice->neighborY,
+ parameterDevice->neighborZ,
+ parameterDevice->numberOfNodes,
+ parameterDevice->isEvenTimestep,
+ vf::lbm::dir::DIR_P00,
+ parameterDevice->outflowPressureCorrectionFactor);
+ getLastCudaError("QPressZeroRhoOutflowDev27 execution failed");
+}
+//////////////////////////////////////////////////////////////////////////
void QInflowScaleByPressDev27(LBMSimulationParameter* parameterDevice, QforBoundaryConditions* boundaryCondition)
{
dim3 grid = vf::cuda::getCudaGrid( parameterDevice->numberofthreads, boundaryCondition->numberOfBCnodes);
@@ -3841,22 +2959,9 @@ void QPressDevOld27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QPressDeviceOld27<<< gridQ, threads >>> ( rhoBC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QPressDeviceOld27<<< grid.grid, grid.threads >>> ( rhoBC,
DD,
k_Q,
k_N,
@@ -3867,7 +2972,7 @@ void QPressDevOld27( unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QPressDeviceOld27 execution failed");
+ getLastCudaError("QPressDeviceOld27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QPressDevIncompNEQ27(LBMSimulationParameter* parameterDevice, QforBoundaryConditions* boundaryCondition)
@@ -3941,22 +3046,9 @@ void QPressDevZero27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QPressDeviceZero27<<< gridQ, threads >>> (DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QPressDeviceZero27<<< grid.grid, grid.threads >>> (DD,
k_Q,
numberOfBCnodes,
neighborX,
@@ -3964,7 +3056,7 @@ void QPressDevZero27(unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QPressDeviceOld27 execution failed");
+ getLastCudaError("QPressDeviceOld27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
void QPressDevFake27( unsigned int numberOfThreads,
@@ -3980,22 +3072,10 @@ void QPressDevFake27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QPressDeviceFake27<<< gridQ, threads >>> (rhoBC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+
+ QPressDeviceFake27<<< grid.grid, grid.threads >>> (rhoBC,
DD,
k_Q,
k_N,
@@ -4040,22 +3120,9 @@ void QPressDev27_IntBB( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- QPressDevice27_IntBB<<< gridQ, threads >>> (rho,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ QPressDevice27_IntBB<<< grid.grid, grid.threads >>> (rho,
DD,
k_Q,
QQ,
@@ -4066,7 +3133,7 @@ void QPressDev27_IntBB( unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("QPressDevice27_IntBB execution failed");
+ getLastCudaError("QPressDevice27_IntBB execution failed");
}
// TODO: https://git.rz.tu-bs.de/irmb/VirtualFluids_dev/-/issues/29
//////////////////////////////////////////////////////////////////////////
@@ -4087,22 +3154,9 @@ void PressSchlaffer27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- PressSchlaff27<<< gridQ, threads >>>( rhoBC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ PressSchlaff27<<< grid.grid, grid.threads >>>( rhoBC,
DD,
vx0,
vy0,
@@ -4117,7 +3171,7 @@ void PressSchlaffer27(unsigned int numberOfThreads,
neighborZ,
size_Mat,
isEvenTimestep);
- getLastCudaError("PressSchlaff27 execution failed");
+ getLastCudaError("PressSchlaff27 execution failed");
}
// TODO: https://git.rz.tu-bs.de/irmb/VirtualFluids_dev/-/issues/29
//////////////////////////////////////////////////////////////////////////
@@ -4136,22 +3190,9 @@ void VelSchlaffer27( unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- VelSchlaff27<<< gridQ, threads >>>( t,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ VelSchlaff27<<< grid.grid, grid.threads >>>( t,
DD,
vz0,
deltaVz0,
@@ -4167,7 +3208,150 @@ void VelSchlaffer27( unsigned int numberOfThreads,
getLastCudaError("VelSchlaff27 execution failed");
}
//////////////////////////////////////////////////////////////////////////
-void PropVelo( unsigned int numberOfThreads,
+void QPrecursorDevCompZeroPress(LBMSimulationParameter* parameterDevice,
+ QforPrecursorBoundaryConditions* boundaryCondition,
+ real timeRatio,
+ real velocityRatio)
+{
+
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(parameterDevice->numberofthreads, boundaryCondition->numberOfBCnodes);
+
+ QPrecursorDeviceCompZeroPress<<< grid.grid, grid.threads >>>(boundaryCondition->k,
+ boundaryCondition->numberOfBCnodes,
+ boundaryCondition->numberOfPrecursorNodes,
+ boundaryCondition->sizeQ,
+ parameterDevice->omega,
+ parameterDevice->distributions.f[0],
+ boundaryCondition->q27[0],
+ parameterDevice->neighborX,
+ parameterDevice->neighborY,
+ parameterDevice->neighborZ,
+ boundaryCondition->planeNeighbor0PP,
+ boundaryCondition->planeNeighbor0PM,
+ boundaryCondition->planeNeighbor0MP,
+ boundaryCondition->planeNeighbor0MM,
+ boundaryCondition->weights0PP,
+ boundaryCondition->weights0PM,
+ boundaryCondition->weights0MP,
+ boundaryCondition->weights0MM,
+ boundaryCondition->last,
+ boundaryCondition->current,
+ boundaryCondition->velocityX,
+ boundaryCondition->velocityY,
+ boundaryCondition->velocityZ,
+ timeRatio,
+ velocityRatio,
+ parameterDevice->numberOfNodes,
+ parameterDevice->isEvenTimestep);
+ getLastCudaError("QPrecursorDeviceCompZeroPress execution failed");
+
+}
+//////////////////////////////////////////////////////////////////////////
+void PrecursorDevEQ27( LBMSimulationParameter* parameterDevice,
+ QforPrecursorBoundaryConditions* boundaryCondition,
+ real timeRatio,
+ real velocityRatio)
+{
+
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(parameterDevice->numberofthreads, boundaryCondition->numberOfBCnodes);
+
+ PrecursorDeviceEQ27<<< grid.grid, grid.threads >>>(boundaryCondition->k,
+ boundaryCondition->numberOfBCnodes,
+ boundaryCondition->numberOfPrecursorNodes,
+ parameterDevice->omega,
+ parameterDevice->distributions.f[0],
+ parameterDevice->neighborX,
+ parameterDevice->neighborX,
+ parameterDevice->neighborX,
+ boundaryCondition->planeNeighbor0PP,
+ boundaryCondition->planeNeighbor0PM,
+ boundaryCondition->planeNeighbor0MP,
+ boundaryCondition->planeNeighbor0MM,
+ boundaryCondition->weights0PP,
+ boundaryCondition->weights0PM,
+ boundaryCondition->weights0MP,
+ boundaryCondition->weights0MM,
+ boundaryCondition->last,
+ boundaryCondition->current,
+ boundaryCondition->velocityX,
+ boundaryCondition->velocityY,
+ boundaryCondition->velocityZ,
+ timeRatio,
+ velocityRatio,
+ parameterDevice->numberOfNodes,
+ parameterDevice->isEvenTimestep);
+ getLastCudaError("PrecursorDeviceEQ27 execution failed");
+
+}
+//////////////////////////////////////////////////////////////////////////
+void PrecursorDevDistributions( LBMSimulationParameter* parameterDevice,
+ QforPrecursorBoundaryConditions* boundaryCondition,
+ real timeRatio,
+ real velocityRatio)
+{
+
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(parameterDevice->numberofthreads, boundaryCondition->numberOfBCnodes);
+
+ PrecursorDeviceDistributions<<< grid.grid, grid.threads >>>(boundaryCondition->k,
+ boundaryCondition->numberOfBCnodes,
+ boundaryCondition->numberOfPrecursorNodes,
+ parameterDevice->distributions.f[0],
+ parameterDevice->neighborX,
+ parameterDevice->neighborY,
+ parameterDevice->neighborZ,
+ boundaryCondition->planeNeighbor0PP,
+ boundaryCondition->planeNeighbor0PM,
+ boundaryCondition->planeNeighbor0MP,
+ boundaryCondition->planeNeighbor0MM,
+ boundaryCondition->weights0PP,
+ boundaryCondition->weights0PM,
+ boundaryCondition->weights0MP,
+ boundaryCondition->weights0MM,
+ boundaryCondition->last,
+ boundaryCondition->current,
+ timeRatio,
+ parameterDevice->numberOfNodes,
+ parameterDevice->isEvenTimestep);
+ getLastCudaError("QPrecursorDeviceCompZeroPress execution failed");
+
+}
+
+//////////////////////////////////////////////////////////////////////////
+void QPrecursorDevDistributions( LBMSimulationParameter* parameterDevice,
+ QforPrecursorBoundaryConditions* boundaryCondition,
+ real timeRatio,
+ real velocityRatio)
+{
+
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(parameterDevice->numberofthreads, boundaryCondition->numberOfBCnodes);
+
+ QPrecursorDeviceDistributions<<< grid.grid, grid.threads >>>(boundaryCondition->k,
+ boundaryCondition->q27[0],
+ boundaryCondition->sizeQ,
+ boundaryCondition->numberOfBCnodes,
+ boundaryCondition->numberOfPrecursorNodes,
+ parameterDevice->distributions.f[0],
+ parameterDevice->neighborX,
+ parameterDevice->neighborY,
+ parameterDevice->neighborZ,
+ boundaryCondition->planeNeighbor0PP,
+ boundaryCondition->planeNeighbor0PM,
+ boundaryCondition->planeNeighbor0MP,
+ boundaryCondition->planeNeighbor0MM,
+ boundaryCondition->weights0PP,
+ boundaryCondition->weights0PM,
+ boundaryCondition->weights0MP,
+ boundaryCondition->weights0MM,
+ boundaryCondition->last,
+ boundaryCondition->current,
+ timeRatio,
+ parameterDevice->numberOfNodes,
+ parameterDevice->isEvenTimestep);
+ getLastCudaError("QPrecursorDeviceCompZeroPress execution failed");
+
+}
+//////////////////////////////////////////////////////////////////////////
+extern "C" void PropVelo( unsigned int numberOfThreads,
unsigned int* neighborX,
unsigned int* neighborY,
unsigned int* neighborZ,
@@ -4182,22 +3366,9 @@ void PropVelo( unsigned int numberOfThreads,
real* DD,
bool EvenOrOdd)
{
- int Grid = (size_Prop / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- PropellerBC<<< grid, threads >>>(neighborX,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Prop);
+
+ PropellerBC<<< grid.grid, grid.threads >>>(neighborX,
neighborY,
neighborZ,
rho,
@@ -4236,22 +3407,9 @@ void ScaleCF27( real* DC,
unsigned int nyF,
unsigned int numberOfThreads)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF27<<< gridINT_CF, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -4299,22 +3457,9 @@ void ScaleCFEff27(real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCFEff27<<< gridINT_CF, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCFEff27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -4363,22 +3508,9 @@ void ScaleCFLast27(real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCFLast27<<< gridINT_CF, threads >>> (DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCFLast27<<< grid.grid, grid.threads >>> (DC,
DF,
neighborCX,
neighborCY,
@@ -4427,22 +3559,9 @@ void ScaleCFpress27( real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCFpress27<<< gridINT_CF, threads >>>(DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCFpress27<<< grid.grid, grid.threads >>>(DC,
DF,
neighborCX,
neighborCY,
@@ -4491,22 +3610,9 @@ void ScaleCF_Fix_27( real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_Fix_27<<< gridINT_CF, threads >>>(DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_Fix_27<<< grid.grid, grid.threads >>>(DC,
DF,
neighborCX,
neighborCY,
@@ -4555,22 +3661,9 @@ void ScaleCF_Fix_comp_27( real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_Fix_comp_27<<< gridINT_CF, threads >>>( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_Fix_comp_27<<< grid.grid, grid.threads >>>( DC,
DF,
neighborCX,
neighborCY,
@@ -4620,22 +3713,9 @@ void ScaleCF_0817_comp_27(real* DC,
OffCF offCF,
CUstream_st *stream)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_0817_comp_27<<< gridINT_CF, threads, 0, stream >>>( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_0817_comp_27<<< grid.grid, grid.threads, 0, stream >>>( DC,
DF,
neighborCX,
neighborCY,
@@ -4685,22 +3765,9 @@ void ScaleCF_comp_D3Q27F3_2018(real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_comp_D3Q27F3_2018 <<< gridINT_CF, threads >>>(DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_comp_D3Q27F3_2018 <<< grid.grid, grid.threads >>>(DC,
DF,
G6,
neighborCX,
@@ -4752,22 +3819,9 @@ void ScaleCF_comp_D3Q27F3(real* DC,
OffCF offCF,
CUstream_st *stream)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_comp_D3Q27F3 <<< gridINT_CF, threads, 0, stream >>>( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_comp_D3Q27F3 <<< grid.grid, grid.threads, 0, stream >>>( DC,
DF,
G6,
neighborCX,
@@ -4817,22 +3871,9 @@ void ScaleCF_staggered_time_comp_27( real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_staggered_time_comp_27<<< gridINT_CF, threads >>>( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_staggered_time_comp_27<<< grid.grid, grid.threads >>>( DC,
DF,
neighborCX,
neighborCY,
@@ -4911,6 +3952,7 @@ void ScaleCF_compressible(LBMSimulationParameter * parameterDeviceC, LBMSimulati
parameterDeviceC->omega,
parameterDeviceF->omega,
offsetCF);
+
getLastCudaError("scaleCF_compressible execution failed");
}
@@ -4940,22 +3982,9 @@ void ScaleCF_RhoSq_3rdMom_comp_27(real* DC,
OffCF offCF,
CUstream_st *stream)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_RhoSq_3rdMom_comp_27<<< gridINT_CF, threads, 0, stream >>>( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_RhoSq_3rdMom_comp_27<<< grid.grid, grid.threads, 0, stream >>>( DC,
DF,
neighborCX,
neighborCY,
@@ -5005,22 +4034,9 @@ void ScaleCF_AA2016_comp_27(real* DC,
OffCF offCF,
CUstream_st *stream)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_AA2016_comp_27<<< gridINT_CF, threads, 0, stream >>>(DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_AA2016_comp_27<<< grid.grid, grid.threads, 0, stream >>>(DC,
DF,
neighborCX,
neighborCY,
@@ -5069,22 +4085,9 @@ void ScaleCF_NSPress_27( real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCF_NSPress_27<<< gridINT_CF, threads >>>(DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCF_NSPress_27<<< grid.grid, grid.threads >>>(DC,
DF,
neighborCX,
neighborCY,
@@ -5130,22 +4133,9 @@ void ScaleCFThSMG7( real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCFThSMG7<<< gridINT_CF, threads >>> (DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCFThSMG7<<< grid.grid, grid.threads >>> (DC,
DF,
DD7C,
DD7F,
@@ -5187,22 +4177,9 @@ void ScaleCFThS7( real* DC,
real diffusivity_fine,
unsigned int numberOfThreads)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCFThS7<<< gridINT_CF, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCFThS7<<< grid.grid, grid.threads >>> ( DC,
DF,
DD7C,
DD7F,
@@ -5244,22 +4221,9 @@ void ScaleCFThS27( real* DC,
unsigned int numberOfThreads,
OffCF offCF)
{
- int Grid = (kCF / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_CF(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleCFThS27<<< gridINT_CF, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kCF);
+
+ scaleCFThS27<<< grid.grid, grid.threads >>> ( DC,
DF,
DD27C,
DD27F,
@@ -5304,22 +4268,10 @@ void ScaleFC27( real* DC,
unsigned int nyF,
unsigned int numberOfThreads)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC27<<< gridINT_FC, threads >>> ( DC,
+
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -5367,22 +4319,9 @@ void ScaleFCEff27(real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFCEff27<<< gridINT_FC, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFCEff27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -5431,22 +4370,9 @@ void ScaleFCLast27(real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFCLast27<<< gridINT_FC, threads >>> (DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFCLast27<<< grid.grid, grid.threads >>> (DC,
DF,
neighborCX,
neighborCY,
@@ -5495,22 +4421,9 @@ void ScaleFCpress27(real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFCpress27<<< gridINT_FC, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFCpress27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -5559,22 +4472,9 @@ void ScaleFC_Fix_27(real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_Fix_27<<< gridINT_FC, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_Fix_27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -5623,22 +4523,9 @@ void ScaleFC_Fix_comp_27( real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_Fix_comp_27<<< gridINT_FC, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_Fix_comp_27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -5688,22 +4575,9 @@ void ScaleFC_0817_comp_27( real* DC,
OffFC offFC,
CUstream_st *stream)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_0817_comp_27<<< gridINT_FC, threads, 0, stream >>> (DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_0817_comp_27<<< grid.grid, grid.threads, 0, stream >>> (DC,
DF,
neighborCX,
neighborCY,
@@ -5753,22 +4627,9 @@ void ScaleFC_comp_D3Q27F3_2018( real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_comp_D3Q27F3_2018 <<< gridINT_FC, threads >>> (DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_comp_D3Q27F3_2018 <<< grid.grid, grid.threads >>> (DC,
DF,
G6,
neighborCX,
@@ -5820,22 +4681,9 @@ void ScaleFC_comp_D3Q27F3( real* DC,
OffFC offFC,
CUstream_st *stream)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_comp_D3Q27F3 <<< gridINT_FC, threads, 0, stream >>> (DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_comp_D3Q27F3 <<< grid.grid, grid.threads, 0, stream >>> (DC,
DF,
G6,
neighborCX,
@@ -5885,22 +4733,9 @@ void ScaleFC_staggered_time_comp_27( real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_staggered_time_comp_27<<< gridINT_FC, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_staggered_time_comp_27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -6007,22 +4842,9 @@ void ScaleFC_RhoSq_3rdMom_comp_27( real* DC,
OffFC offFC,
CUstream_st *stream)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_RhoSq_3rdMom_comp_27<<< gridINT_FC, threads, 0, stream >>>(DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_RhoSq_3rdMom_comp_27<<< grid.grid, grid.threads, 0, stream >>>(DC,
DF,
neighborCX,
neighborCY,
@@ -6072,22 +4894,9 @@ void ScaleFC_AA2016_comp_27( real* DC,
OffFC offFC,
CUstream_st *stream)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_AA2016_comp_27<<< gridINT_FC, threads, 0, stream >>>(DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_AA2016_comp_27<<< grid.grid, grid.threads, 0, stream >>>(DC,
DF,
neighborCX,
neighborCY,
@@ -6136,22 +4945,9 @@ void ScaleFC_NSPress_27(real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFC_NSPress_27<<< gridINT_FC, threads >>> ( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFC_NSPress_27<<< grid.grid, grid.threads >>> ( DC,
DF,
neighborCX,
neighborCY,
@@ -6197,22 +4993,9 @@ void ScaleFCThSMG7(real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFCThSMG7<<< gridINT_FC, threads >>>( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFCThSMG7<<< grid.grid, grid.threads >>>( DC,
DF,
DD7C,
DD7F,
@@ -6254,22 +5037,9 @@ void ScaleFCThS7( real* DC,
real diffusivity_coarse,
unsigned int numberOfThreads)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFCThS7<<< gridINT_FC, threads >>>(DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFCThS7<<< grid.grid, grid.threads >>>(DC,
DF,
DD7C,
DD7F,
@@ -6311,22 +5081,9 @@ void ScaleFCThS27( real* DC,
unsigned int numberOfThreads,
OffFC offFC)
{
- int Grid = (kFC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridINT_FC(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- scaleFCThS27<<< gridINT_FC, threads >>>( DC,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, kFC);
+
+ scaleFCThS27<<< grid.grid, grid.threads >>>( DC,
DF,
DD27C,
DD27F,
@@ -6362,22 +5119,9 @@ void DragLiftPostD27(real* DD,
bool isEvenTimestep,
unsigned int numberOfThreads)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- DragLiftPost27<<< grid, threads >>>(DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ DragLiftPost27<<< grid.grid, grid.threads >>>(DD,
k_Q,
QQ,
numberOfBCnodes,
@@ -6406,22 +5150,9 @@ void DragLiftPreD27( real* DD,
bool isEvenTimestep,
unsigned int numberOfThreads)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- DragLiftPre27<<< grid, threads >>>( DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ DragLiftPre27<<< grid.grid, grid.threads >>>( DD,
k_Q,
QQ,
numberOfBCnodes,
@@ -6447,22 +5178,9 @@ void CalcCPtop27(real* DD,
bool isEvenTimestep,
unsigned int numberOfThreads)
{
- int Grid = (nonCp / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- CalcCP27<<< grid, threads >>>(DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, nonCp);
+
+ CalcCP27<<< grid.grid, grid.threads >>>(DD,
cpIndex,
nonCp,
cpPress,
@@ -6485,22 +5203,9 @@ void CalcCPbottom27( real* DD,
bool isEvenTimestep,
unsigned int numberOfThreads)
{
- int Grid = (nonCp / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- CalcCP27<<< grid, threads >>>(DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, nonCp);
+
+ CalcCP27<<< grid.grid, grid.threads >>>(DD,
cpIndex,
nonCp,
cpPress,
@@ -6524,22 +5229,9 @@ void GetSendFsPreDev27(real* DD,
unsigned int numberOfThreads,
cudaStream_t stream)
{
- int Grid = (buffmax / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- getSendFsPre27<<< grid, threads, 0, stream >>>(DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, buffmax);
+
+ getSendFsPre27<<< grid.grid, grid.threads, 0, stream >>>(DD,
bufferFs,
sendIndex,
buffmax,
@@ -6563,22 +5255,9 @@ void GetSendFsPostDev27(real* DD,
unsigned int numberOfThreads,
cudaStream_t stream)
{
- int Grid = (buffmax / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- getSendFsPost27<<< grid, threads, 0, stream >>>(DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, buffmax);
+
+ getSendFsPost27<<< grid.grid, grid.threads, 0, stream >>>(DD,
bufferFs,
sendIndex,
buffmax,
@@ -6602,22 +5281,9 @@ void SetRecvFsPreDev27(real* DD,
unsigned int numberOfThreads,
cudaStream_t stream)
{
- int Grid = (buffmax / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- setRecvFsPre27<<< grid, threads, 0, stream >>>(DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, buffmax);
+
+ setRecvFsPre27<<< grid.grid, grid.threads, 0, stream >>>(DD,
bufferFs,
recvIndex,
buffmax,
@@ -6641,22 +5307,9 @@ void SetRecvFsPostDev27(real* DD,
unsigned int numberOfThreads,
cudaStream_t stream)
{
- int Grid = (buffmax / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- setRecvFsPost27<<< grid, threads, 0, stream >>>(DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, buffmax);
+
+ setRecvFsPost27<<< grid.grid, grid.threads, 0, stream >>>(DD,
bufferFs,
recvIndex,
buffmax,
@@ -6680,22 +5333,9 @@ void getSendGsDevF3(
bool isEvenTimestep,
unsigned int numberOfThreads)
{
- int Grid = (buffmax / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1);
-
- getSendGsF3 <<< grid, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, buffmax);
+
+ getSendGsF3 <<< grid.grid, grid.threads >>> (
G6,
bufferGs,
sendIndex,
@@ -6720,22 +5360,9 @@ void setRecvGsDevF3(
bool isEvenTimestep,
unsigned int numberOfThreads)
{
- int Grid = (buffmax / numberOfThreads) + 1;
- int Grid1, Grid2;
- if (Grid > 512)
- {
- Grid1 = 512;
- Grid2 = (Grid / Grid1) + 1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 grid(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1);
-
- setRecvGsF3 <<< grid, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, buffmax);
+
+ setRecvGsF3 <<< grid.grid, grid.threads >>> (
G6,
bufferGs,
recvIndex,
@@ -6763,22 +5390,9 @@ void WallFuncDev27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (numberOfBCnodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- WallFunction27<<< gridQ, threads >>> (
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfBCnodes);
+
+ WallFunction27<<< grid.grid, grid.threads >>> (
vx,
vy,
vz,
@@ -6814,22 +5428,9 @@ void SetOutputWallVelocitySP27(unsigned int numberOfThreads,
real* DD,
bool isEvenTimestep)
{
- int Grid = (numberOfWallNodes / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- LBSetOutputWallVelocitySP27<<< gridQ, threads >>> ( vxD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfWallNodes);
+
+ LBSetOutputWallVelocitySP27<<< grid.grid, grid.threads >>> ( vxD,
vyD,
vzD,
vxWall,
@@ -6862,22 +5463,9 @@ void GetVelotoForce27(unsigned int numberOfThreads,
unsigned int size_Mat,
bool isEvenTimestep)
{
- int Grid = (nonAtBC / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- GetVeloforForcing27<<< gridQ, threads >>> (DD,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, nonAtBC);
+
+ GetVeloforForcing27<<< grid.grid, grid.threads >>> (DD,
bcIndex,
nonAtBC,
Vx,
@@ -6911,27 +5499,14 @@ void InitParticlesDevice(real* coordX,
unsigned int* neighborY,
unsigned int* neighborZ,
unsigned int* neighborWSB,
- int level,
+ int level,
unsigned int numberOfParticles,
unsigned int size_Mat,
unsigned int numberOfThreads)
{
- int Grid = (numberOfParticles / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- InitParticles<<< gridQ, threads >>> (coordX,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfParticles);
+
+ InitParticles<<< grid.grid, grid.threads >>> (coordX,
coordY,
coordZ,
coordParticleXlocal,
@@ -6986,22 +5561,9 @@ void MoveParticlesDevice(real* coordX,
unsigned int numberOfThreads,
bool isEvenTimestep)
{
- int Grid = (numberOfParticles / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- MoveParticles<<< gridQ, threads >>> (coordX,
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, numberOfParticles);
+
+ MoveParticles<<< grid.grid, grid.threads >>> (coordX,
coordY,
coordZ,
coordParticleXlocal,
@@ -7035,22 +5597,8 @@ void initRandomDevice(curandState* state,
unsigned int size_Mat,
unsigned int numberOfThreads)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- initRandom<<< gridQ, threads >>> (state);
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+ initRandom<<< grid.grid, grid.threads >>> (state);
getLastCudaError("initRandom execution failed");
}
//////////////////////////////////////////////////////////////////////////
@@ -7059,22 +5607,8 @@ void generateRandomValuesDevice( curandState* state,
real* randArray,
unsigned int numberOfThreads)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- generateRandomValues<<< gridQ, threads >>> (state,randArray);
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+ generateRandomValues<<< grid.grid, grid.threads >>> (state,randArray);
getLastCudaError("generateRandomValues execution failed");
}
//////////////////////////////////////////////////////////////////////////
@@ -7097,22 +5631,8 @@ void CalcTurbulenceIntensityDevice(
bool isEvenTimestep,
uint numberOfThreads)
{
- int Grid = (size_Mat / numberOfThreads)+1;
- int Grid1, Grid2;
- if (Grid>512)
- {
- Grid1 = 512;
- Grid2 = (Grid/Grid1)+1;
- }
- else
- {
- Grid1 = 1;
- Grid2 = Grid;
- }
- dim3 gridQ(Grid1, Grid2);
- dim3 threads(numberOfThreads, 1, 1 );
-
- CalcTurbulenceIntensity<<<gridQ, threads>>>(
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(numberOfThreads, size_Mat);
+ CalcTurbulenceIntensity<<<grid.grid, grid.threads>>>(
vxx,
vyy,
vzz,
diff --git a/src/gpu/VirtualFluids_GPU/GPU/PrecursorBCs27.cu b/src/gpu/VirtualFluids_GPU/GPU/PrecursorBCs27.cu
new file mode 100644
index 0000000000000000000000000000000000000000..78b190e37ebaa395c89aae3b47cd4cc4f3147306
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/GPU/PrecursorBCs27.cu
@@ -0,0 +1,967 @@
+#include "LBM/LB.h"
+#include <lbm/constants/NumericConstants.h>
+#include <lbm/constants/D3Q27.h>
+#include <lbm/MacroscopicQuantities.h>
+
+#include "VirtualFluids_GPU/Kernel/Utilities/DistributionHelper.cuh"
+#include "VirtualFluids_GPU/GPU/KernelUtilities.h"
+
+using namespace vf::lbm::constant;
+using namespace vf::lbm::dir;
+
+__global__ void QPrecursorDeviceCompZeroPress( int* subgridDistanceIndices,
+ int numberOfBCnodes,
+ int numberOfPrecursorNodes,
+ int sizeQ,
+ real omega,
+ real* distributions,
+ real* subgridDistances,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* neighbors0PP,
+ uint* neighbors0PM,
+ uint* neighbors0MP,
+ uint* neighbors0MM,
+ real* weights0PP,
+ real* weights0PM,
+ real* weights0MP,
+ real* weights0MM,
+ real* vLast,
+ real* vCurrent,
+ real velocityX,
+ real velocityY,
+ real velocityZ,
+ real timeRatio,
+ real velocityRatio,
+ unsigned long long numberOfLBnodes,
+ bool isEvenTimestep)
+{
+ const unsigned k = vf::gpu::getNodeIndex();
+
+ if(k>=numberOfBCnodes) return;
+
+ ////////////////////////////////////////////////////////////////////////////////
+ // interpolation of velocity
+ real vxLastInterpd, vyLastInterpd, vzLastInterpd;
+ real vxNextInterpd, vyNextInterpd, vzNextInterpd;
+
+ uint kNeighbor0PP = neighbors0PP[k];
+ real d0PP = weights0PP[k];
+
+ real* vxLast = vLast;
+ real* vyLast = &vLast[numberOfPrecursorNodes];
+ real* vzLast = &vLast[2*numberOfPrecursorNodes];
+
+ real* vxCurrent = vCurrent;
+ real* vyCurrent = &vCurrent[numberOfPrecursorNodes];
+ real* vzCurrent = &vCurrent[2*numberOfPrecursorNodes];
+
+ if(d0PP < 1e6)
+ {
+ uint kNeighbor0PM = neighbors0PM[k];
+ uint kNeighbor0MP = neighbors0MP[k];
+ uint kNeighbor0MM = neighbors0MM[k];
+
+ real d0PM = weights0PM[k];
+ real d0MP = weights0MP[k];
+ real d0MM = weights0MM[k];
+
+ real invWeightSum = 1.f/(d0PP+d0PM+d0MP+d0MM);
+
+ vxLastInterpd = (vxLast[kNeighbor0PP]*d0PP + vxLast[kNeighbor0PM]*d0PM + vxLast[kNeighbor0MP]*d0MP + vxLast[kNeighbor0MM]*d0MM)*invWeightSum;
+ vyLastInterpd = (vyLast[kNeighbor0PP]*d0PP + vyLast[kNeighbor0PM]*d0PM + vyLast[kNeighbor0MP]*d0MP + vyLast[kNeighbor0MM]*d0MM)*invWeightSum;
+ vzLastInterpd = (vzLast[kNeighbor0PP]*d0PP + vzLast[kNeighbor0PM]*d0PM + vzLast[kNeighbor0MP]*d0MP + vzLast[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ vxNextInterpd = (vxCurrent[kNeighbor0PP]*d0PP + vxCurrent[kNeighbor0PM]*d0PM + vxCurrent[kNeighbor0MP]*d0MP + vxCurrent[kNeighbor0MM]*d0MM)*invWeightSum;
+ vyNextInterpd = (vyCurrent[kNeighbor0PP]*d0PP + vyCurrent[kNeighbor0PM]*d0PM + vyCurrent[kNeighbor0MP]*d0MP + vyCurrent[kNeighbor0MM]*d0MM)*invWeightSum;
+ vzNextInterpd = (vzCurrent[kNeighbor0PP]*d0PP + vzCurrent[kNeighbor0PM]*d0PM + vzCurrent[kNeighbor0MP]*d0MP + vzCurrent[kNeighbor0MM]*d0MM)*invWeightSum;
+ }
+ else
+ {
+ vxLastInterpd = vxLast[kNeighbor0PP];
+ vyLastInterpd = vyLast[kNeighbor0PP];
+ vzLastInterpd = vzLast[kNeighbor0PP];
+
+ vxNextInterpd = vxCurrent[kNeighbor0PP];
+ vyNextInterpd = vyCurrent[kNeighbor0PP];
+ vzNextInterpd = vzCurrent[kNeighbor0PP];
+ }
+
+ // if(k==16300)s printf("%f %f %f\n", vxLastInterpd, vyLastInterpd, vzLastInterpd);
+ real VeloX = (velocityX + (1.f-timeRatio)*vxLastInterpd + timeRatio*vxNextInterpd)/velocityRatio;
+ real VeloY = (velocityY + (1.f-timeRatio)*vyLastInterpd + timeRatio*vyNextInterpd)/velocityRatio;
+ real VeloZ = (velocityZ + (1.f-timeRatio)*vzLastInterpd + timeRatio*vzNextInterpd)/velocityRatio;
+ // From here on just a copy of QVelDeviceCompZeroPress
+ ////////////////////////////////////////////////////////////////////////////////
+
+ Distributions27 dist;
+ getPointersToDistributions(dist, distributions, numberOfLBnodes, isEvenTimestep);
+
+ unsigned int KQK = subgridDistanceIndices[k];
+ unsigned int k000= KQK;
+ unsigned int kP00 = KQK;
+ unsigned int kM00 = neighborX[KQK];
+ unsigned int k0P0 = KQK;
+ unsigned int k0M0 = neighborY[KQK];
+ unsigned int k00P = KQK;
+ unsigned int k00M = neighborZ[KQK];
+ unsigned int kMM0 = neighborY[kM00];
+ unsigned int kPP0 = KQK;
+ unsigned int kPM0 = k0M0;
+ unsigned int kMP0 = kM00;
+ unsigned int kM0M = neighborZ[kM00];
+ unsigned int kP0P = KQK;
+ unsigned int kP0M = k00M;
+ unsigned int kM0P = kM00;
+ unsigned int k0PP = KQK;
+ unsigned int k0MM = neighborZ[k0M0];
+ unsigned int k0PM = k00M;
+ unsigned int k0MP = k0M0;
+ unsigned int kPMP = k0M0;
+ unsigned int kMPM = kM0M;
+ unsigned int kMPP = kM00;
+ unsigned int kPMM = k0MM;
+ unsigned int kMMP = kMM0;
+ unsigned int kPPM = k00M;
+ unsigned int kPPP = KQK;
+ unsigned int kMMM = neighborZ[kMM0];
+
+ ////////////////////////////////////////////////////////////////////////////////
+ //! - Set local distributions
+ //!
+ real f_M00 = (dist.f[DIR_P00])[kP00];
+ real f_P00 = (dist.f[DIR_M00])[kM00];
+ real f_0M0 = (dist.f[DIR_0P0])[k0P0];
+ real f_0P0 = (dist.f[DIR_0M0])[k0M0];
+ real f_00M = (dist.f[DIR_00P])[k00P];
+ real f_00P = (dist.f[DIR_00M])[k00M];
+ real f_MM0 = (dist.f[DIR_PP0])[kPP0];
+ real f_PP0 = (dist.f[DIR_MM0])[kMM0];
+ real f_MP0 = (dist.f[DIR_PM0])[kPM0];
+ real f_PM0 = (dist.f[DIR_MP0])[kMP0];
+ real f_M0M = (dist.f[DIR_P0P])[kP0P];
+ real f_P0P = (dist.f[DIR_M0M])[kM0M];
+ real f_M0P = (dist.f[DIR_P0M])[kP0M];
+ real f_P0M = (dist.f[DIR_M0P])[kM0P];
+ real f_0MM = (dist.f[DIR_0PP])[k0PP];
+ real f_0PP = (dist.f[DIR_0MM])[k0MM];
+ real f_0MP = (dist.f[DIR_0PM])[k0PM];
+ real f_0PM = (dist.f[DIR_0MP])[k0MP];
+ real f_MMM = (dist.f[DIR_PPP])[kPPP];
+ real f_PPM = (dist.f[DIR_MMP])[kMMP];
+ real f_MPM = (dist.f[DIR_PMP])[kPMP];
+ real f_PMM = (dist.f[DIR_MPP])[kMPP];
+ real f_MMP = (dist.f[DIR_PPM])[kPPM];
+ real f_PPP = (dist.f[DIR_MMM])[kMMM];
+ real f_MPP = (dist.f[DIR_PMM])[kPMM];
+ real f_PMP = (dist.f[DIR_MPM])[kMPM];
+
+ SubgridDistances27 subgridD;
+ getPointersToSubgridDistances(subgridD, subgridDistances, numberOfBCnodes);
+
+ ////////////////////////////////////////////////////////////////////////////////
+ real drho = f_PMP + f_MPP + f_PPP + f_MMP + f_PMM + f_MPM + f_PPM + f_MMM +
+ f_0PM + f_0PP + f_0MP + f_0MM + f_P0M + f_M0P + f_P0P + f_M0M + f_PM0 + f_MP0 + f_PP0 + f_MM0 +
+ f_00P + f_00M + f_0P0 + f_0M0 + f_P00 + f_M00 + ((dist.f[DIR_000])[k000]);
+
+ real vx1 = (((f_PMP - f_MPM) - (f_MPP - f_PMM)) + ((f_PPP - f_MMM) - (f_MMP - f_PPM)) +
+ ((f_P0M - f_M0P) + (f_P0P - f_M0M)) + ((f_PM0 - f_MP0) + (f_PP0 - f_MM0)) +
+ (f_P00 - f_M00)) / (c1o1 + drho);
+
+
+ real vx2 = ((-(f_PMP - f_MPM) + (f_MPP - f_PMM)) + ((f_PPP - f_MMM) - (f_MMP - f_PPM)) +
+ ((f_0PM - f_0MP) + (f_0PP - f_0MM)) + (-(f_PM0 - f_MP0) + (f_PP0 - f_MM0)) +
+ (f_0P0 - f_0M0)) / (c1o1 + drho);
+
+ real vx3 = (((f_PMP - f_MPM) + (f_MPP - f_PMM)) + ((f_PPP - f_MMM) + (f_MMP - f_PPM)) +
+ (-(f_0PM - f_0MP) + (f_0PP - f_0MM)) + ((f_P0P - f_M0M) - (f_P0M - f_M0P)) +
+ (f_00P - f_00M)) / (c1o1 + drho);
+
+
+ // if(k==16383 || k==0) printf("k %d kQ %d drho = %f u %f v %f w %f\n",k, KQK, drho, vx1, vx2, vx3);
+ real cu_sq=c3o2*(vx1*vx1+vx2*vx2+vx3*vx3) * (c1o1 + drho);
+ //////////////////////////////////////////////////////////////////////////
+
+
+ ////////////////////////////////////////////////////////////////////////////////
+ //! - Update distributions with subgrid distance (q) between zero and one
+ real feq, q, velocityLB, velocityBC;
+ q = (subgridD.q[DIR_P00])[k];
+ if (q>=c0o1 && q<=c1o1) // only update distribution for q between zero and one
+ {
+ velocityLB = vx1;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c2o27);
+ velocityBC = VeloX;
+ (dist.f[DIR_M00])[kM00] = getInterpolatedDistributionForVeloWithPressureBC(q, f_P00, f_M00, feq, omega, drho, velocityBC, c2o27);
+ }
+
+ q = (subgridD.q[DIR_M00])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c2o27);
+ velocityBC = -VeloX;
+ (dist.f[DIR_P00])[kP00] = getInterpolatedDistributionForVeloWithPressureBC(q, f_M00, f_P00, feq, omega, drho, velocityBC, c2o27);
+ }
+
+ q = (subgridD.q[DIR_0P0])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx2;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c2o27);
+ velocityBC = VeloY;
+ (dist.f[DIR_0M0])[DIR_0M0] = getInterpolatedDistributionForVeloWithPressureBC(q, f_0P0, f_0M0, feq, omega, drho, velocityBC, c2o27);
+ }
+
+ q = (subgridD.q[DIR_0M0])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx2;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c2o27);
+ velocityBC = -VeloY;
+ (dist.f[DIR_0P0])[k0P0] = getInterpolatedDistributionForVeloWithPressureBC(q, f_0M0, f_0P0, feq, omega, drho, velocityBC, c2o27);
+ }
+
+ q = (subgridD.q[DIR_00P])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c2o27);
+ velocityBC = VeloZ;
+ (dist.f[DIR_00M])[k00M] = getInterpolatedDistributionForVeloWithPressureBC(q, f_00P, f_00M, feq, omega, drho, velocityBC, c2o27);
+ }
+
+ q = (subgridD.q[DIR_00M])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c2o27);
+ velocityBC = -VeloZ;
+ (dist.f[DIR_00P])[k00P] = getInterpolatedDistributionForVeloWithPressureBC(q, f_00M, f_00P, feq, omega, drho, velocityBC, c2o27);
+ }
+
+ q = (subgridD.q[DIR_PP0])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx1 + vx2;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = VeloX + VeloY;
+ (dist.f[DIR_MM0])[kMM0] = getInterpolatedDistributionForVeloWithPressureBC(q, f_PP0, f_MM0, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_MM0])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1 - vx2;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = -VeloX - VeloY;
+ (dist.f[DIR_PP0])[kPP0] = getInterpolatedDistributionForVeloWithPressureBC(q, f_MM0, f_PP0, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_PM0])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx1 - vx2;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = VeloX - VeloY;
+ (dist.f[DIR_MP0])[kMP0] = getInterpolatedDistributionForVeloWithPressureBC(q, f_PM0, f_MP0, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_MP0])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1 + vx2;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = -VeloX + VeloY;
+ (dist.f[DIR_PM0])[kPM0] = getInterpolatedDistributionForVeloWithPressureBC(q, f_MP0, f_PM0, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_P0P])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx1 + vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = VeloX + VeloZ;
+ (dist.f[DIR_M0M])[kM0M] = getInterpolatedDistributionForVeloWithPressureBC(q, f_P0P, f_M0M, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_M0M])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1 - vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = -VeloX - VeloZ;
+ (dist.f[DIR_P0P])[kP0P] = getInterpolatedDistributionForVeloWithPressureBC(q, f_M0M, f_P0P, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_P0M])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx1 - vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = VeloX - VeloZ;
+ (dist.f[DIR_M0P])[kM0P] = getInterpolatedDistributionForVeloWithPressureBC(q, f_P0M, f_M0P, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_M0P])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1 + vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = -VeloX + VeloZ;
+ (dist.f[DIR_P0M])[kP0M] = getInterpolatedDistributionForVeloWithPressureBC(q, f_M0P, f_P0M, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_0PP])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx2 + vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = VeloY + VeloZ;
+ (dist.f[DIR_0MM])[k0MM] = getInterpolatedDistributionForVeloWithPressureBC(q, f_0PP, f_0MM, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_0MM])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx2 - vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = -VeloY - VeloZ;
+ (dist.f[DIR_0PP])[k0PP] = getInterpolatedDistributionForVeloWithPressureBC(q, f_0MM, f_0PP, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_0PM])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx2 - vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = VeloY - VeloZ;
+ (dist.f[DIR_0MP])[k0MP] = getInterpolatedDistributionForVeloWithPressureBC(q, f_0PM, f_0PP, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_0MP])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx2 + vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o54);
+ velocityBC = -VeloY + VeloZ;
+ (dist.f[DIR_0PM])[k0PM] = getInterpolatedDistributionForVeloWithPressureBC(q, f_0PP, f_0PM, feq, omega, drho, velocityBC, c1o54);
+ }
+
+ q = (subgridD.q[DIR_PPP])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx1 + vx2 + vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o216);
+ velocityBC = VeloX + VeloY + VeloZ;
+ (dist.f[DIR_MMM])[kMMM] = getInterpolatedDistributionForVeloWithPressureBC(q, f_PPP, f_MMM, feq, omega, drho, velocityBC, c1o216);
+ }
+
+ q = (subgridD.q[DIR_MMM])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1 - vx2 - vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o216);
+ velocityBC = -VeloX - VeloY - VeloZ;
+ (dist.f[DIR_PPP])[kPPP] = getInterpolatedDistributionForVeloWithPressureBC(q, f_MMM, f_PPP, feq, omega, drho, velocityBC, c1o216);
+ }
+
+ q = (subgridD.q[DIR_PPM])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx1 + vx2 - vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o216);
+ velocityBC = VeloX + VeloY - VeloZ;
+ (dist.f[DIR_MMP])[kMMP] = getInterpolatedDistributionForVeloWithPressureBC(q, f_PPM, f_MMP, feq, omega, drho, velocityBC, c1o216);
+ }
+
+ q = (subgridD.q[DIR_MMP])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1 - vx2 + vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o216);
+ velocityBC = -VeloX - VeloY + VeloZ;
+ (dist.f[DIR_PPM])[kPPM] = getInterpolatedDistributionForVeloWithPressureBC(q, f_MMP, f_PPM, feq, omega, drho, velocityBC, c1o216);
+ }
+
+ q = (subgridD.q[DIR_PMP])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx1 - vx2 + vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o216);
+ velocityBC = VeloX - VeloY + VeloZ;
+ (dist.f[DIR_MPM])[kMPM] = getInterpolatedDistributionForVeloWithPressureBC(q, f_PMP, f_MPM, feq, omega, drho, velocityBC, c1o216);
+ }
+
+ q = (subgridD.q[DIR_MPM])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1 + vx2 - vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o216);
+ velocityBC = -VeloX + VeloY - VeloZ;
+ (dist.f[DIR_PMP])[kPMP] = getInterpolatedDistributionForVeloWithPressureBC(q, f_MPM, f_PMP, feq, omega, drho, velocityBC, c1o216);
+ }
+
+ q = (subgridD.q[DIR_PMM])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = vx1 - vx2 - vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o216);
+ velocityBC = VeloX - VeloY - VeloZ;
+ (dist.f[DIR_MPP])[kMPP] = getInterpolatedDistributionForVeloWithPressureBC(q, f_PMM, f_MPP, feq, omega, drho, velocityBC, c1o216);
+ }
+
+ q = (subgridD.q[DIR_MPP])[k];
+ if (q>=c0o1 && q<=c1o1)
+ {
+ velocityLB = -vx1 + vx2 + vx3;
+ feq = getEquilibriumForBC(drho, velocityLB, cu_sq, c1o216);
+ velocityBC = -VeloX + VeloY + VeloZ;
+ (dist.f[DIR_PMM])[kPMM] = getInterpolatedDistributionForVeloWithPressureBC(q, f_MPP, f_PMM, feq, omega, drho, velocityBC, c1o216);
+ }
+}
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+
+
+__global__ void PrecursorDeviceEQ27( int* subgridDistanceIndices,
+ int numberOfBCnodes,
+ int numberOfPrecursorNodes,
+ real omega,
+ real* distributions,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* neighbors0PP,
+ uint* neighbors0PM,
+ uint* neighbors0MP,
+ uint* neighbors0MM,
+ real* weights0PP,
+ real* weights0PM,
+ real* weights0MP,
+ real* weights0MM,
+ real* vLast,
+ real* vCurrent,
+ real velocityX,
+ real velocityY,
+ real velocityZ,
+ real timeRatio,
+ real velocityRatio,
+ unsigned long long numberOfLBnodes,
+ bool isEvenTimestep)
+{
+ const unsigned k = vf::gpu::getNodeIndex();
+
+ if(k>=numberOfBCnodes) return;
+
+ ////////////////////////////////////////////////////////////////////////////////
+ // interpolation of velocity
+ real vxLastInterpd, vyLastInterpd, vzLastInterpd;
+ real vxNextInterpd, vyNextInterpd, vzNextInterpd;
+
+ uint kNeighbor0PP = neighbors0PP[k];
+ real d0PP = weights0PP[k];
+
+ real* vxLast = vLast;
+ real* vyLast = &vLast[numberOfPrecursorNodes];
+ real* vzLast = &vLast[2*numberOfPrecursorNodes];
+
+ real* vxCurrent = vCurrent;
+ real* vyCurrent = &vCurrent[numberOfPrecursorNodes];
+ real* vzCurrent = &vCurrent[2*numberOfPrecursorNodes];
+
+ if(d0PP < 1e6)
+ {
+ uint kNeighbor0PM = neighbors0PM[k];
+ uint kNeighbor0MP = neighbors0MP[k];
+ uint kNeighbor0MM = neighbors0MM[k];
+
+ real d0PM = weights0PM[k];
+ real d0MP = weights0MP[k];
+ real d0MM = weights0MM[k];
+
+ real invWeightSum = 1.f/(d0PP+d0PM+d0MP+d0MM);
+
+ vxLastInterpd = (vxLast[kNeighbor0PP]*d0PP + vxLast[kNeighbor0PM]*d0PM + vxLast[kNeighbor0MP]*d0MP + vxLast[kNeighbor0MM]*d0MM)*invWeightSum;
+ vyLastInterpd = (vyLast[kNeighbor0PP]*d0PP + vyLast[kNeighbor0PM]*d0PM + vyLast[kNeighbor0MP]*d0MP + vyLast[kNeighbor0MM]*d0MM)*invWeightSum;
+ vzLastInterpd = (vzLast[kNeighbor0PP]*d0PP + vzLast[kNeighbor0PM]*d0PM + vzLast[kNeighbor0MP]*d0MP + vzLast[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ vxNextInterpd = (vxCurrent[kNeighbor0PP]*d0PP + vxCurrent[kNeighbor0PM]*d0PM + vxCurrent[kNeighbor0MP]*d0MP + vxCurrent[kNeighbor0MM]*d0MM)*invWeightSum;
+ vyNextInterpd = (vyCurrent[kNeighbor0PP]*d0PP + vyCurrent[kNeighbor0PM]*d0PM + vyCurrent[kNeighbor0MP]*d0MP + vyCurrent[kNeighbor0MM]*d0MM)*invWeightSum;
+ vzNextInterpd = (vzCurrent[kNeighbor0PP]*d0PP + vzCurrent[kNeighbor0PM]*d0PM + vzCurrent[kNeighbor0MP]*d0MP + vzCurrent[kNeighbor0MM]*d0MM)*invWeightSum;
+ }
+ else
+ {
+ vxLastInterpd = vxLast[kNeighbor0PP];
+ vyLastInterpd = vyLast[kNeighbor0PP];
+ vzLastInterpd = vzLast[kNeighbor0PP];
+
+ vxNextInterpd = vxCurrent[kNeighbor0PP];
+ vyNextInterpd = vyCurrent[kNeighbor0PP];
+ vzNextInterpd = vzCurrent[kNeighbor0PP];
+ }
+
+ // if(k==16300) printf("%f %f %f\n", vxLastInterpd, vyLastInterpd, vzLastInterpd);
+ real VeloX = (velocityX + (1.f-timeRatio)*vxLastInterpd + timeRatio*vxNextInterpd)/velocityRatio;
+ real VeloY = (velocityY + (1.f-timeRatio)*vyLastInterpd + timeRatio*vyNextInterpd)/velocityRatio;
+ real VeloZ = (velocityZ + (1.f-timeRatio)*vzLastInterpd + timeRatio*vzNextInterpd)/velocityRatio;
+ // From here on just a copy of QVelDeviceCompZeroPress
+ ////////////////////////////////////////////////////////////////////////////////
+
+ Distributions27 dist;
+ getPointersToDistributions(dist, distributions, numberOfLBnodes, !isEvenTimestep);
+
+ unsigned int KQK = subgridDistanceIndices[k]; //QK
+ unsigned int k000 = KQK; //000
+ unsigned int kP00 = KQK; //P00
+ unsigned int kM00 = neighborX[KQK]; //M00
+ unsigned int k0P0 = KQK; //n
+ unsigned int k0M0 = neighborY[KQK]; //s
+ unsigned int k00P = KQK; //t
+ unsigned int k00M = neighborZ[KQK]; //b
+ unsigned int kMM0 = neighborY[kM00]; //sw
+ unsigned int kPP0 = KQK; //ne
+ unsigned int kPM0 = k0M0; //se
+ unsigned int kMP0 = kM00; //nw
+ unsigned int kM0M = neighborZ[kM00]; //bw
+ unsigned int kP0P = KQK; //te
+ unsigned int kP0M = k00M; //be
+ unsigned int k0PP = KQK; //tn
+ unsigned int k0MM = neighborZ[k0M0]; //bs
+ unsigned int kM0P = kM00; //tw
+ unsigned int k0PM = k00M; //bn
+ unsigned int k0MP = k0M0; //ts
+ unsigned int kPMP = k0M0; //tse
+ unsigned int kMPM = kM0M; //bnw
+ unsigned int kMPP = kM00; //tnw
+ unsigned int kPMM = k0MM; //bse
+ unsigned int kMMP = kMM0; //tsw
+ unsigned int kPPM = k00M; //bne
+ unsigned int kPPP = KQK; //tne
+ unsigned int kMMM = neighborZ[kMM0]; //bsw
+
+ //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ // based on BGK Plus Comp
+ //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ real f_M00 = (dist.f[DIR_P00])[kP00];
+ real f_P00 = (dist.f[DIR_M00])[kM00];
+ real f_0M0 = (dist.f[DIR_0P0])[k0P0];
+ real f_0P0 = (dist.f[DIR_0M0])[k0M0];
+ real f_00M = (dist.f[DIR_00P])[k00P];
+ real f_00P = (dist.f[DIR_00M])[k00M];
+ real f_MM0 = (dist.f[DIR_PP0])[kPP0];
+ real f_PP0 = (dist.f[DIR_MM0])[kMM0];
+ real f_MP0 = (dist.f[DIR_PM0])[kPM0];
+ real f_PM0 = (dist.f[DIR_MP0])[kMP0];
+ real f_M0M = (dist.f[DIR_P0P])[kP0P];
+ real f_P0P = (dist.f[DIR_M0M])[kM0M];
+ real f_M0P = (dist.f[DIR_P0M])[kP0M];
+ real f_P0M = (dist.f[DIR_M0P])[kM0P];
+ real f_0MM = (dist.f[DIR_0PP])[k0PP];
+ real f_0PP = (dist.f[DIR_0MM])[k0MM];
+ real f_0PM = (dist.f[DIR_0MP])[k0MP];
+ real f_0MP = (dist.f[DIR_0PM])[k0PM];
+ real f_000 = (dist.f[DIR_000])[k000];
+ real f_MMM = (dist.f[DIR_PPP])[kPPP];
+ real f_PPM = (dist.f[DIR_MMP])[kMMP];
+ real f_MPM = (dist.f[DIR_PMP])[kPMP];
+ real f_PMM = (dist.f[DIR_MPP])[kMPP];
+ real f_MMP = (dist.f[DIR_PPM])[kPPM];
+ real f_PPP = (dist.f[DIR_MMM])[kMMM];
+ real f_MPP = (dist.f[DIR_PMM])[kPMM];
+ real f_PMP = (dist.f[DIR_MPM])[kMPM];
+
+ ////////////////////////////////////////////////////////////////////////////////
+ //! - Set macroscopic quantities
+ //!
+ real drho = c0o1;
+
+ real vx1 = VeloX;
+
+ real vx2 = VeloY;
+
+ real vx3 = VeloZ;
+
+ real cusq = c3o2 * (vx1 * vx1 + vx2 * vx2 + vx3 * vx3);
+
+ ////////////////////////////////////////////////////////////////////////////////
+ f_000 = c8o27* (drho-(drho+c1o1)*cusq);
+ f_P00 = c2o27* (drho+(drho+c1o1)*(c3o1*( vx1 )+c9o2*( vx1 )*( vx1 )-cusq));
+ f_M00 = c2o27* (drho+(drho+c1o1)*(c3o1*(-vx1 )+c9o2*(-vx1 )*(-vx1 )-cusq));
+ f_0P0 = c2o27* (drho+(drho+c1o1)*(c3o1*( vx2 )+c9o2*( vx2 )*( vx2 )-cusq));
+ f_0M0 = c2o27* (drho+(drho+c1o1)*(c3o1*( -vx2 )+c9o2*( -vx2 )*( -vx2 )-cusq));
+ f_00P = c2o27* (drho+(drho+c1o1)*(c3o1*( vx3)+c9o2*( vx3)*( vx3)-cusq));
+ f_00M = c2o27* (drho+(drho+c1o1)*(c3o1*( -vx3)+c9o2*( -vx3)*( -vx3)-cusq));
+ f_PP0 = c1o54* (drho+(drho+c1o1)*(c3o1*( vx1+vx2 )+c9o2*( vx1+vx2 )*( vx1+vx2 )-cusq));
+ f_MM0 = c1o54* (drho+(drho+c1o1)*(c3o1*(-vx1-vx2 )+c9o2*(-vx1-vx2 )*(-vx1-vx2 )-cusq));
+ f_PM0 = c1o54* (drho+(drho+c1o1)*(c3o1*( vx1-vx2 )+c9o2*( vx1-vx2 )*( vx1-vx2 )-cusq));
+ f_MP0 = c1o54* (drho+(drho+c1o1)*(c3o1*(-vx1+vx2 )+c9o2*(-vx1+vx2 )*(-vx1+vx2 )-cusq));
+ f_P0P = c1o54* (drho+(drho+c1o1)*(c3o1*( vx1 +vx3)+c9o2*( vx1 +vx3)*( vx1 +vx3)-cusq));
+ f_M0M = c1o54* (drho+(drho+c1o1)*(c3o1*(-vx1 -vx3)+c9o2*(-vx1 -vx3)*(-vx1 -vx3)-cusq));
+ f_P0M = c1o54* (drho+(drho+c1o1)*(c3o1*( vx1 -vx3)+c9o2*( vx1 -vx3)*( vx1 -vx3)-cusq));
+ f_M0P = c1o54* (drho+(drho+c1o1)*(c3o1*(-vx1 +vx3)+c9o2*(-vx1 +vx3)*(-vx1 +vx3)-cusq));
+ f_0PP = c1o54* (drho+(drho+c1o1)*(c3o1*( vx2+vx3)+c9o2*( vx2+vx3)*( vx2+vx3)-cusq));
+ f_0MM = c1o54* (drho+(drho+c1o1)*(c3o1*( -vx2-vx3)+c9o2*( -vx2-vx3)*( -vx2-vx3)-cusq));
+ f_0PM = c1o54* (drho+(drho+c1o1)*(c3o1*( vx2-vx3)+c9o2*( vx2-vx3)*( vx2-vx3)-cusq));
+ f_0MP = c1o54* (drho+(drho+c1o1)*(c3o1*( -vx2+vx3)+c9o2*( -vx2+vx3)*( -vx2+vx3)-cusq));
+ f_PPP = c1o216*(drho+(drho+c1o1)*(c3o1*( vx1+vx2+vx3)+c9o2*( vx1+vx2+vx3)*( vx1+vx2+vx3)-cusq));
+ f_MMM = c1o216*(drho+(drho+c1o1)*(c3o1*(-vx1-vx2-vx3)+c9o2*(-vx1-vx2-vx3)*(-vx1-vx2-vx3)-cusq));
+ f_PPM = c1o216*(drho+(drho+c1o1)*(c3o1*( vx1+vx2-vx3)+c9o2*( vx1+vx2-vx3)*( vx1+vx2-vx3)-cusq));
+ f_MMP = c1o216*(drho+(drho+c1o1)*(c3o1*(-vx1-vx2+vx3)+c9o2*(-vx1-vx2+vx3)*(-vx1-vx2+vx3)-cusq));
+ f_PMP = c1o216*(drho+(drho+c1o1)*(c3o1*( vx1-vx2+vx3)+c9o2*( vx1-vx2+vx3)*( vx1-vx2+vx3)-cusq));
+ f_MPM = c1o216*(drho+(drho+c1o1)*(c3o1*(-vx1+vx2-vx3)+c9o2*(-vx1+vx2-vx3)*(-vx1+vx2-vx3)-cusq));
+ f_PMM = c1o216*(drho+(drho+c1o1)*(c3o1*( vx1-vx2-vx3)+c9o2*( vx1-vx2-vx3)*( vx1-vx2-vx3)-cusq));
+ f_MPP = c1o216*(drho+(drho+c1o1)*(c3o1*(-vx1+vx2+vx3)+c9o2*(-vx1+vx2+vx3)*(-vx1+vx2+vx3)-cusq));
+
+ ////////////////////////////////////////////////////////////////////////////////
+ //! write the new distributions to the bc nodes
+ //!
+ (dist.f[DIR_P00])[kP00] = f_M00;
+ (dist.f[DIR_PP0])[kPP0] = f_MM0;
+ (dist.f[DIR_P0M])[kP0M] = f_M0P;
+ (dist.f[DIR_PM0])[kPM0] = f_MP0;
+ (dist.f[DIR_PMP])[kPMP] = f_MPM;
+ (dist.f[DIR_P0P])[kP0P] = f_M0M;
+ (dist.f[DIR_PPM])[kPPM] = f_MMP;
+ (dist.f[DIR_PPP])[kPPP] = f_MMM;
+ (dist.f[DIR_PMM])[kPMM] = f_MPP;
+
+ (dist.f[DIR_M00])[kM00] = f_P00;
+ (dist.f[DIR_MM0])[kMM0] = f_PP0;
+ (dist.f[DIR_M0M])[kM0M] = f_P0P;
+ (dist.f[DIR_MP0])[kMP0] = f_PM0;
+ (dist.f[DIR_M0P])[kM0P] = f_P0M;
+ (dist.f[DIR_MMM])[kMMM] = f_PPP;
+ (dist.f[DIR_MMP])[kMMP] = f_PPM;
+ (dist.f[DIR_MPP])[kMPP] = f_PMM;
+ (dist.f[DIR_MPM])[kMPM] = f_PMP;
+
+ (dist.f[DIR_0P0])[k0P0] = f_0M0;
+ (dist.f[DIR_0M0])[k0M0] = f_0P0;
+ (dist.f[DIR_00P])[k00P] = f_00M;
+ (dist.f[DIR_00M])[k00M] = f_00P;
+ (dist.f[DIR_0PP])[k0PP] = f_0MM;
+ (dist.f[DIR_0MM])[k0MM] = f_0PP;
+ (dist.f[DIR_0PM])[k0PM] = f_0MP;
+ (dist.f[DIR_0MP])[k0MP] = f_0PM;
+ (dist.f[DIR_000])[k000] = f_000;
+}
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+
+
+__global__ void PrecursorDeviceDistributions( int* subgridDistanceIndices,
+ int numberOfBCnodes,
+ int numberOfPrecursorNodes,
+ real* distributions,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* neighbors0PP,
+ uint* neighbors0PM,
+ uint* neighbors0MP,
+ uint* neighbors0MM,
+ real* weights0PP,
+ real* weights0PM,
+ real* weights0MP,
+ real* weights0MM,
+ real* fsLast,
+ real* fsNext,
+ real timeRatio,
+ unsigned long long numberOfLBnodes,
+ bool isEvenTimestep)
+{
+ const unsigned k = vf::gpu::getNodeIndex();
+
+ if(k>=numberOfBCnodes) return;
+
+ uint kNeighbor0PP = neighbors0PP[k];
+ real d0PP = weights0PP[k];
+
+ real f0LastInterp, f1LastInterp, f2LastInterp, f3LastInterp, f4LastInterp, f5LastInterp, f6LastInterp, f7LastInterp, f8LastInterp;
+ real f0NextInterp, f1NextInterp, f2NextInterp, f3NextInterp, f4NextInterp, f5NextInterp, f6NextInterp, f7NextInterp, f8NextInterp;
+
+ real* f0Last = fsLast;
+ real* f1Last = &fsLast[ numberOfPrecursorNodes];
+ real* f2Last = &fsLast[2*numberOfPrecursorNodes];
+ real* f3Last = &fsLast[3*numberOfPrecursorNodes];
+ real* f4Last = &fsLast[4*numberOfPrecursorNodes];
+ real* f5Last = &fsLast[5*numberOfPrecursorNodes];
+ real* f6Last = &fsLast[6*numberOfPrecursorNodes];
+ real* f7Last = &fsLast[7*numberOfPrecursorNodes];
+ real* f8Last = &fsLast[8*numberOfPrecursorNodes];
+
+ real* f0Next = fsNext;
+ real* f1Next = &fsNext[ numberOfPrecursorNodes];
+ real* f2Next = &fsNext[2*numberOfPrecursorNodes];
+ real* f3Next = &fsNext[3*numberOfPrecursorNodes];
+ real* f4Next = &fsNext[4*numberOfPrecursorNodes];
+ real* f5Next = &fsNext[5*numberOfPrecursorNodes];
+ real* f6Next = &fsNext[6*numberOfPrecursorNodes];
+ real* f7Next = &fsNext[7*numberOfPrecursorNodes];
+ real* f8Next = &fsNext[8*numberOfPrecursorNodes];
+
+
+ if(d0PP<1e6)
+ {
+ uint kNeighbor0PM = neighbors0PM[k];
+ uint kNeighbor0MP = neighbors0MP[k];
+ uint kNeighbor0MM = neighbors0MM[k];
+
+ real d0PM = weights0PM[k];
+ real d0MP = weights0MP[k];
+ real d0MM = weights0MM[k];
+
+ real invWeightSum = 1.f/(d0PP+d0PM+d0MP+d0MM);
+
+ f0LastInterp = (f0Last[kNeighbor0PP]*d0PP + f0Last[kNeighbor0PM]*d0PM + f0Last[kNeighbor0MP]*d0MP + f0Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f0NextInterp = (f0Next[kNeighbor0PP]*d0PP + f0Next[kNeighbor0PM]*d0PM + f0Next[kNeighbor0MP]*d0MP + f0Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f1LastInterp = (f1Last[kNeighbor0PP]*d0PP + f1Last[kNeighbor0PM]*d0PM + f1Last[kNeighbor0MP]*d0MP + f1Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f1NextInterp = (f1Next[kNeighbor0PP]*d0PP + f1Next[kNeighbor0PM]*d0PM + f1Next[kNeighbor0MP]*d0MP + f1Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f2LastInterp = (f2Last[kNeighbor0PP]*d0PP + f2Last[kNeighbor0PM]*d0PM + f2Last[kNeighbor0MP]*d0MP + f2Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f2NextInterp = (f2Next[kNeighbor0PP]*d0PP + f2Next[kNeighbor0PM]*d0PM + f2Next[kNeighbor0MP]*d0MP + f2Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f3LastInterp = (f3Last[kNeighbor0PP]*d0PP + f3Last[kNeighbor0PM]*d0PM + f3Last[kNeighbor0MP]*d0MP + f3Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f3NextInterp = (f3Next[kNeighbor0PP]*d0PP + f3Next[kNeighbor0PM]*d0PM + f3Next[kNeighbor0MP]*d0MP + f3Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f4LastInterp = (f4Last[kNeighbor0PP]*d0PP + f4Last[kNeighbor0PM]*d0PM + f4Last[kNeighbor0MP]*d0MP + f4Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f4NextInterp = (f4Next[kNeighbor0PP]*d0PP + f4Next[kNeighbor0PM]*d0PM + f4Next[kNeighbor0MP]*d0MP + f4Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f5LastInterp = (f5Last[kNeighbor0PP]*d0PP + f5Last[kNeighbor0PM]*d0PM + f5Last[kNeighbor0MP]*d0MP + f5Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f5NextInterp = (f5Next[kNeighbor0PP]*d0PP + f5Next[kNeighbor0PM]*d0PM + f5Next[kNeighbor0MP]*d0MP + f5Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f6LastInterp = (f6Last[kNeighbor0PP]*d0PP + f6Last[kNeighbor0PM]*d0PM + f6Last[kNeighbor0MP]*d0MP + f6Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f6NextInterp = (f6Next[kNeighbor0PP]*d0PP + f6Next[kNeighbor0PM]*d0PM + f6Next[kNeighbor0MP]*d0MP + f6Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f7LastInterp = (f7Last[kNeighbor0PP]*d0PP + f7Last[kNeighbor0PM]*d0PM + f7Last[kNeighbor0MP]*d0MP + f7Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f7NextInterp = (f7Next[kNeighbor0PP]*d0PP + f7Next[kNeighbor0PM]*d0PM + f7Next[kNeighbor0MP]*d0MP + f7Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f8LastInterp = (f8Last[kNeighbor0PP]*d0PP + f8Last[kNeighbor0PM]*d0PM + f8Last[kNeighbor0MP]*d0MP + f8Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f8NextInterp = (f8Next[kNeighbor0PP]*d0PP + f8Next[kNeighbor0PM]*d0PM + f8Next[kNeighbor0MP]*d0MP + f8Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ } else {
+ f0LastInterp = f0Last[kNeighbor0PP];
+ f1LastInterp = f1Last[kNeighbor0PP];
+ f2LastInterp = f2Last[kNeighbor0PP];
+ f3LastInterp = f3Last[kNeighbor0PP];
+ f4LastInterp = f4Last[kNeighbor0PP];
+ f5LastInterp = f5Last[kNeighbor0PP];
+ f6LastInterp = f6Last[kNeighbor0PP];
+ f7LastInterp = f7Last[kNeighbor0PP];
+ f8LastInterp = f8Last[kNeighbor0PP];
+
+ f0NextInterp = f0Next[kNeighbor0PP];
+ f1NextInterp = f1Next[kNeighbor0PP];
+ f2NextInterp = f2Next[kNeighbor0PP];
+ f3NextInterp = f3Next[kNeighbor0PP];
+ f4NextInterp = f4Next[kNeighbor0PP];
+ f5NextInterp = f5Next[kNeighbor0PP];
+ f6NextInterp = f6Next[kNeighbor0PP];
+ f7NextInterp = f7Next[kNeighbor0PP];
+ f8NextInterp = f8Next[kNeighbor0PP];
+ }
+ Distributions27 dist;
+ getPointersToDistributions(dist, distributions, numberOfLBnodes, !isEvenTimestep);
+
+ unsigned int KQK = subgridDistanceIndices[k];
+ // unsigned int k000= KQK;
+ unsigned int kP00 = KQK;
+ // unsigned int kM00 = neighborX[KQK];
+ // unsigned int k0P0 = KQK;
+ unsigned int k0M0 = neighborY[KQK];
+ // unsigned int k00P = KQK;
+ unsigned int k00M = neighborZ[KQK];
+ // unsigned int kMM0 = neighborY[kM00];
+ unsigned int kPP0 = KQK;
+ unsigned int kPM0 = k0M0;
+ // unsigned int kMP0 = kM00;
+ // unsigned int kM0M = neighborZ[kM00];
+ unsigned int kP0P = KQK;
+ unsigned int kP0M = k00M;
+ // unsigned int kM0P = kM00;
+ unsigned int k0MM = neighborZ[k0M0];
+ // unsigned int k0PM = k00M;
+ // unsigned int k0MP = k0M0;
+ unsigned int kPMP = k0M0;
+ // unsigned int kMPM = kM0M;
+ // unsigned int kMPP = kM00;
+ unsigned int kPMM = k0MM;
+ // unsigned int kMMP = kMM0;
+ unsigned int kPPM = k00M;
+ unsigned int kPPP = KQK;
+ // unsigned int kMMM = neighborZ[kMM0];
+
+ dist.f[DIR_P00][kP00] = f0LastInterp*(1.f-timeRatio) + f0NextInterp*timeRatio;
+ dist.f[DIR_PP0][kPP0] = f1LastInterp*(1.f-timeRatio) + f1NextInterp*timeRatio;
+ dist.f[DIR_PM0][kPM0] = f2LastInterp*(1.f-timeRatio) + f2NextInterp*timeRatio;
+ dist.f[DIR_P0P][kP0P] = f3LastInterp*(1.f-timeRatio) + f3NextInterp*timeRatio;
+ dist.f[DIR_P0M][kP0M] = f4LastInterp*(1.f-timeRatio) + f4NextInterp*timeRatio;
+ dist.f[DIR_PPP][kPPP] = f5LastInterp*(1.f-timeRatio) + f5NextInterp*timeRatio;
+ dist.f[DIR_PMP][kPMP] = f6LastInterp*(1.f-timeRatio) + f6NextInterp*timeRatio;
+ dist.f[DIR_PPM][kPPM] = f7LastInterp*(1.f-timeRatio) + f7NextInterp*timeRatio;
+ dist.f[DIR_PMM][kPMM] = f8LastInterp*(1.f-timeRatio) + f8NextInterp*timeRatio;
+}
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+
+//NOTE: Has not been tested after bug fix!
+__global__ void QPrecursorDeviceDistributions( int* subgridDistanceIndices,
+ real* subgridDistances,
+ int sizeQ,
+ int numberOfBCnodes,
+ int numberOfPrecursorNodes,
+ real* distributions,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* neighbors0PP,
+ uint* neighbors0PM,
+ uint* neighbors0MP,
+ uint* neighbors0MM,
+ real* weights0PP,
+ real* weights0PM,
+ real* weights0MP,
+ real* weights0MM,
+ real* fsLast,
+ real* fsNext,
+ real timeRatio,
+ unsigned long long numberOfLBnodes,
+ bool isEvenTimestep)
+{
+ const unsigned k = vf::gpu::getNodeIndex();
+
+ if(k>=numberOfBCnodes) return;
+
+ uint kNeighbor0PP = neighbors0PP[k];
+ real d0PP = weights0PP[k];
+
+ real f0LastInterp, f1LastInterp, f2LastInterp, f3LastInterp, f4LastInterp, f5LastInterp, f6LastInterp, f7LastInterp, f8LastInterp;
+ real f0NextInterp, f1NextInterp, f2NextInterp, f3NextInterp, f4NextInterp, f5NextInterp, f6NextInterp, f7NextInterp, f8NextInterp;
+
+ real* f0Last = fsLast;
+ real* f1Last = &fsLast[ numberOfPrecursorNodes];
+ real* f2Last = &fsLast[2*numberOfPrecursorNodes];
+ real* f3Last = &fsLast[3*numberOfPrecursorNodes];
+ real* f4Last = &fsLast[4*numberOfPrecursorNodes];
+ real* f5Last = &fsLast[5*numberOfPrecursorNodes];
+ real* f6Last = &fsLast[6*numberOfPrecursorNodes];
+ real* f7Last = &fsLast[7*numberOfPrecursorNodes];
+ real* f8Last = &fsLast[8*numberOfPrecursorNodes];
+
+ real* f0Next = fsNext;
+ real* f1Next = &fsNext[ numberOfPrecursorNodes];
+ real* f2Next = &fsNext[2*numberOfPrecursorNodes];
+ real* f3Next = &fsNext[3*numberOfPrecursorNodes];
+ real* f4Next = &fsNext[4*numberOfPrecursorNodes];
+ real* f5Next = &fsNext[5*numberOfPrecursorNodes];
+ real* f6Next = &fsNext[6*numberOfPrecursorNodes];
+ real* f7Next = &fsNext[7*numberOfPrecursorNodes];
+ real* f8Next = &fsNext[8*numberOfPrecursorNodes];
+
+
+ if(d0PP<1e6)
+ {
+ uint kNeighbor0PM = neighbors0PM[k];
+ uint kNeighbor0MP = neighbors0MP[k];
+ uint kNeighbor0MM = neighbors0MM[k];
+
+ real d0PM = weights0PM[k];
+ real d0MP = weights0MP[k];
+ real d0MM = weights0MM[k];
+
+ real invWeightSum = 1.f/(d0PP+d0PM+d0MP+d0MM);
+
+ f0LastInterp = (f0Last[kNeighbor0PP]*d0PP + f0Last[kNeighbor0PM]*d0PM + f0Last[kNeighbor0MP]*d0MP + f0Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f0NextInterp = (f0Next[kNeighbor0PP]*d0PP + f0Next[kNeighbor0PM]*d0PM + f0Next[kNeighbor0MP]*d0MP + f0Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f1LastInterp = (f1Last[kNeighbor0PP]*d0PP + f1Last[kNeighbor0PM]*d0PM + f1Last[kNeighbor0MP]*d0MP + f1Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f1NextInterp = (f1Next[kNeighbor0PP]*d0PP + f1Next[kNeighbor0PM]*d0PM + f1Next[kNeighbor0MP]*d0MP + f1Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f2LastInterp = (f2Last[kNeighbor0PP]*d0PP + f2Last[kNeighbor0PM]*d0PM + f2Last[kNeighbor0MP]*d0MP + f2Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f2NextInterp = (f2Next[kNeighbor0PP]*d0PP + f2Next[kNeighbor0PM]*d0PM + f2Next[kNeighbor0MP]*d0MP + f2Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f3LastInterp = (f3Last[kNeighbor0PP]*d0PP + f3Last[kNeighbor0PM]*d0PM + f3Last[kNeighbor0MP]*d0MP + f3Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f3NextInterp = (f3Next[kNeighbor0PP]*d0PP + f3Next[kNeighbor0PM]*d0PM + f3Next[kNeighbor0MP]*d0MP + f3Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f4LastInterp = (f4Last[kNeighbor0PP]*d0PP + f4Last[kNeighbor0PM]*d0PM + f4Last[kNeighbor0MP]*d0MP + f4Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f4NextInterp = (f4Next[kNeighbor0PP]*d0PP + f4Next[kNeighbor0PM]*d0PM + f4Next[kNeighbor0MP]*d0MP + f4Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f5LastInterp = (f5Last[kNeighbor0PP]*d0PP + f5Last[kNeighbor0PM]*d0PM + f5Last[kNeighbor0MP]*d0MP + f5Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f5NextInterp = (f5Next[kNeighbor0PP]*d0PP + f5Next[kNeighbor0PM]*d0PM + f5Next[kNeighbor0MP]*d0MP + f5Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f6LastInterp = (f6Last[kNeighbor0PP]*d0PP + f6Last[kNeighbor0PM]*d0PM + f6Last[kNeighbor0MP]*d0MP + f6Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f6NextInterp = (f6Next[kNeighbor0PP]*d0PP + f6Next[kNeighbor0PM]*d0PM + f6Next[kNeighbor0MP]*d0MP + f6Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f7LastInterp = (f7Last[kNeighbor0PP]*d0PP + f7Last[kNeighbor0PM]*d0PM + f7Last[kNeighbor0MP]*d0MP + f7Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f7NextInterp = (f7Next[kNeighbor0PP]*d0PP + f7Next[kNeighbor0PM]*d0PM + f7Next[kNeighbor0MP]*d0MP + f7Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ f8LastInterp = (f8Last[kNeighbor0PP]*d0PP + f8Last[kNeighbor0PM]*d0PM + f8Last[kNeighbor0MP]*d0MP + f8Last[kNeighbor0MM]*d0MM)*invWeightSum;
+ f8NextInterp = (f8Next[kNeighbor0PP]*d0PP + f8Next[kNeighbor0PM]*d0PM + f8Next[kNeighbor0MP]*d0MP + f8Next[kNeighbor0MM]*d0MM)*invWeightSum;
+
+ } else {
+ f0LastInterp = f0Last[kNeighbor0PP];
+ f1LastInterp = f1Last[kNeighbor0PP];
+ f2LastInterp = f2Last[kNeighbor0PP];
+ f3LastInterp = f3Last[kNeighbor0PP];
+ f4LastInterp = f4Last[kNeighbor0PP];
+ f5LastInterp = f5Last[kNeighbor0PP];
+ f6LastInterp = f6Last[kNeighbor0PP];
+ f7LastInterp = f7Last[kNeighbor0PP];
+ f8LastInterp = f8Last[kNeighbor0PP];
+
+ f0NextInterp = f0Next[kNeighbor0PP];
+ f1NextInterp = f1Next[kNeighbor0PP];
+ f2NextInterp = f2Next[kNeighbor0PP];
+ f3NextInterp = f3Next[kNeighbor0PP];
+ f4NextInterp = f4Next[kNeighbor0PP];
+ f5NextInterp = f5Next[kNeighbor0PP];
+ f6NextInterp = f6Next[kNeighbor0PP];
+ f7NextInterp = f7Next[kNeighbor0PP];
+ f8NextInterp = f8Next[kNeighbor0PP];
+ }
+ Distributions27 dist;
+ getPointersToDistributions(dist, distributions, numberOfLBnodes, !isEvenTimestep);
+
+ unsigned int KQK = subgridDistanceIndices[k];
+ // unsigned int k000= KQK;
+ unsigned int kP00 = KQK;
+ // unsigned int kM00 = neighborX[KQK];
+ // unsigned int k0P0 = KQK;
+ unsigned int k0M0 = neighborY[KQK];
+ // unsigned int k00P = KQK;
+ unsigned int k00M = neighborZ[KQK];
+ // unsigned int kMM0 = neighborY[kM00];
+ unsigned int kPP0 = KQK;
+ unsigned int kPM0 = k0M0;
+ // unsigned int kMP0 = kM00;
+ // unsigned int kM0M = neighborZ[kM00];
+ unsigned int kP0P = KQK;
+ unsigned int kP0M = k00M;
+ // unsigned int kM0P = kM00;
+ unsigned int k0MM = neighborZ[k0M0];
+ // unsigned int k0PM = k00M;
+ // unsigned int k0MP = k0M0;
+ unsigned int kPMP = k0M0;
+ // unsigned int kMPM = kM0M;
+ // unsigned int kMPP = kM00;
+ unsigned int kPMM = k0MM;
+ // unsigned int kMMP = kMM0;
+ unsigned int kPPM = k00M;
+ unsigned int kPPP = KQK;
+ // unsigned int kMMM = neighborZ[kMM0];
+ SubgridDistances27 qs;
+ getPointersToSubgridDistances(qs, subgridDistances, sizeQ);
+
+ real q;
+ q = qs.q[DIR_P00][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_P00][kP00] = f0LastInterp*(1.f-timeRatio) + f0NextInterp*timeRatio;
+ q = qs.q[DIR_PP0][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_PP0][kPP0] = f1LastInterp*(1.f-timeRatio) + f1NextInterp*timeRatio;
+ q = qs.q[DIR_PM0][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_PM0][kPM0] = f2LastInterp*(1.f-timeRatio) + f2NextInterp*timeRatio;
+ q = qs.q[DIR_P0P][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_P0P][kP0P] = f3LastInterp*(1.f-timeRatio) + f3NextInterp*timeRatio;
+ q = qs.q[DIR_P0M][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_P0M][kP0M] = f4LastInterp*(1.f-timeRatio) + f4NextInterp*timeRatio;
+ q = qs.q[DIR_PPP][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_PPP][kPPP] = f5LastInterp*(1.f-timeRatio) + f5NextInterp*timeRatio;
+ q = qs.q[DIR_PMP][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_PMP][kPMP] = f6LastInterp*(1.f-timeRatio) + f6NextInterp*timeRatio;
+ q = qs.q[DIR_PPM][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_PPM][kPPM] = f7LastInterp*(1.f-timeRatio) + f7NextInterp*timeRatio;
+ q = qs.q[DIR_PMM][k]; if(q>= c0o1 && q <= c1o1) dist.f[DIR_PMM][kPMM] = f8LastInterp*(1.f-timeRatio) + f8NextInterp*timeRatio;
+
+}
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
diff --git a/src/gpu/VirtualFluids_GPU/GPU/PressBCs27.cu b/src/gpu/VirtualFluids_GPU/GPU/PressBCs27.cu
index ccb2ce79c63515e59e4f9ae75016f44ced71a170..6a14cebd465a1e79acf14ea019abb34e66d9c85f 100644
--- a/src/gpu/VirtualFluids_GPU/GPU/PressBCs27.cu
+++ b/src/gpu/VirtualFluids_GPU/GPU/PressBCs27.cu
@@ -2,6 +2,9 @@
#include "LBM/LB.h"
#include "lbm/constants/D3Q27.h"
#include "lbm/constants/NumericConstants.h"
+#include "lbm/MacroscopicQuantities.h"
+#include "Kernel/Utilities/DistributionHelper.cuh"
+
#include "KernelUtilities.h"
using namespace vf::lbm::constant;
@@ -2793,12 +2796,14 @@ __global__ void QPressDeviceDirDepBot27( real* rhoBC,
-
-
+__host__ __device__ real computeOutflowDistribution(const real* const &f, const real* const &f1, const int dir, const real cs)
+{
+ return f1[dir] * cs + (c1o1 - cs) * f[dir];
+}
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-__global__ void QPressNoRhoDevice27( real* rhoBC,
- real* DD,
+__global__ void QPressNoRhoDevice27( real* rhoBC,
+ real* distributions,
int* k_Q,
int* k_N,
int numberOfBCnodes,
@@ -2806,483 +2811,419 @@ __global__ void QPressNoRhoDevice27( real* rhoBC,
unsigned int* neighborX,
unsigned int* neighborY,
unsigned int* neighborZ,
- unsigned int size_Mat,
- bool isEvenTimestep)
+ unsigned int numberOfLBnodes,
+ bool isEvenTimestep,
+ int direction)
{
////////////////////////////////////////////////////////////////////////////////
- const unsigned x = threadIdx.x; // Globaler x-Index
- const unsigned y = blockIdx.x; // Globaler y-Index
- const unsigned z = blockIdx.y; // Globaler z-Index
- const unsigned nx = blockDim.x;
- const unsigned ny = gridDim.x;
- const unsigned k = nx*(ny*z + y) + x;
+ const unsigned k = vf::gpu::getNodeIndex();
//////////////////////////////////////////////////////////////////////////
- if(k<numberOfBCnodes)
- {
- ////////////////////////////////////////////////////////////////////////////////
- //index
- unsigned int KQK = k_Q[k];
- //unsigned int kzero= KQK;
- unsigned int ke = KQK;
- unsigned int kw = neighborX[KQK];
- unsigned int kn = KQK;
- unsigned int ks = neighborY[KQK];
- unsigned int kt = KQK;
- unsigned int kb = neighborZ[KQK];
- unsigned int ksw = neighborY[kw];
- unsigned int kne = KQK;
- unsigned int kse = ks;
- unsigned int knw = kw;
- unsigned int kbw = neighborZ[kw];
- unsigned int kte = KQK;
- unsigned int kbe = kb;
- unsigned int ktw = kw;
- unsigned int kbs = neighborZ[ks];
- unsigned int ktn = KQK;
- unsigned int kbn = kb;
- unsigned int kts = ks;
- unsigned int ktse = ks;
- unsigned int kbnw = kbw;
- unsigned int ktnw = kw;
- unsigned int kbse = kbs;
- unsigned int ktsw = ksw;
- unsigned int kbne = kb;
- unsigned int ktne = KQK;
- unsigned int kbsw = neighborZ[ksw];
- ////////////////////////////////////////////////////////////////////////////////
- //index1
- unsigned int K1QK = k_N[k];
- //unsigned int k1zero= K1QK;
- unsigned int k1e = K1QK;
- unsigned int k1w = neighborX[K1QK];
- unsigned int k1n = K1QK;
- unsigned int k1s = neighborY[K1QK];
- unsigned int k1t = K1QK;
- unsigned int k1b = neighborZ[K1QK];
- unsigned int k1sw = neighborY[k1w];
- unsigned int k1ne = K1QK;
- unsigned int k1se = k1s;
- unsigned int k1nw = k1w;
- unsigned int k1bw = neighborZ[k1w];
- unsigned int k1te = K1QK;
- unsigned int k1be = k1b;
- unsigned int k1tw = k1w;
- unsigned int k1bs = neighborZ[k1s];
- unsigned int k1tn = K1QK;
- unsigned int k1bn = k1b;
- unsigned int k1ts = k1s;
- unsigned int k1tse = k1s;
- unsigned int k1bnw = k1bw;
- unsigned int k1tnw = k1w;
- unsigned int k1bse = k1bs;
- unsigned int k1tsw = k1sw;
- unsigned int k1bne = k1b;
- unsigned int k1tne = K1QK;
- unsigned int k1bsw = neighborZ[k1sw];
- ////////////////////////////////////////////////////////////////////////////////
- Distributions27 D;
- if (isEvenTimestep==true)
- {
- D.f[DIR_P00 ] = &DD[DIR_P00 *size_Mat];
- D.f[DIR_M00 ] = &DD[DIR_M00 *size_Mat];
- D.f[DIR_0P0 ] = &DD[DIR_0P0 *size_Mat];
- D.f[DIR_0M0 ] = &DD[DIR_0M0 *size_Mat];
- D.f[DIR_00P ] = &DD[DIR_00P *size_Mat];
- D.f[DIR_00M ] = &DD[DIR_00M *size_Mat];
- D.f[DIR_PP0 ] = &DD[DIR_PP0 *size_Mat];
- D.f[DIR_MM0 ] = &DD[DIR_MM0 *size_Mat];
- D.f[DIR_PM0 ] = &DD[DIR_PM0 *size_Mat];
- D.f[DIR_MP0 ] = &DD[DIR_MP0 *size_Mat];
- D.f[DIR_P0P ] = &DD[DIR_P0P *size_Mat];
- D.f[DIR_M0M ] = &DD[DIR_M0M *size_Mat];
- D.f[DIR_P0M ] = &DD[DIR_P0M *size_Mat];
- D.f[DIR_M0P ] = &DD[DIR_M0P *size_Mat];
- D.f[DIR_0PP ] = &DD[DIR_0PP *size_Mat];
- D.f[DIR_0MM ] = &DD[DIR_0MM *size_Mat];
- D.f[DIR_0PM ] = &DD[DIR_0PM *size_Mat];
- D.f[DIR_0MP ] = &DD[DIR_0MP *size_Mat];
- D.f[DIR_000] = &DD[DIR_000*size_Mat];
- D.f[DIR_PPP ] = &DD[DIR_PPP *size_Mat];
- D.f[DIR_MMP ] = &DD[DIR_MMP *size_Mat];
- D.f[DIR_PMP ] = &DD[DIR_PMP *size_Mat];
- D.f[DIR_MPP ] = &DD[DIR_MPP *size_Mat];
- D.f[DIR_PPM ] = &DD[DIR_PPM *size_Mat];
- D.f[DIR_MMM ] = &DD[DIR_MMM *size_Mat];
- D.f[DIR_PMM ] = &DD[DIR_PMM *size_Mat];
- D.f[DIR_MPM ] = &DD[DIR_MPM *size_Mat];
- }
- else
- {
- D.f[DIR_M00 ] = &DD[DIR_P00 *size_Mat];
- D.f[DIR_P00 ] = &DD[DIR_M00 *size_Mat];
- D.f[DIR_0M0 ] = &DD[DIR_0P0 *size_Mat];
- D.f[DIR_0P0 ] = &DD[DIR_0M0 *size_Mat];
- D.f[DIR_00M ] = &DD[DIR_00P *size_Mat];
- D.f[DIR_00P ] = &DD[DIR_00M *size_Mat];
- D.f[DIR_MM0 ] = &DD[DIR_PP0 *size_Mat];
- D.f[DIR_PP0 ] = &DD[DIR_MM0 *size_Mat];
- D.f[DIR_MP0 ] = &DD[DIR_PM0 *size_Mat];
- D.f[DIR_PM0 ] = &DD[DIR_MP0 *size_Mat];
- D.f[DIR_M0M ] = &DD[DIR_P0P *size_Mat];
- D.f[DIR_P0P ] = &DD[DIR_M0M *size_Mat];
- D.f[DIR_M0P ] = &DD[DIR_P0M *size_Mat];
- D.f[DIR_P0M ] = &DD[DIR_M0P *size_Mat];
- D.f[DIR_0MM ] = &DD[DIR_0PP *size_Mat];
- D.f[DIR_0PP ] = &DD[DIR_0MM *size_Mat];
- D.f[DIR_0MP ] = &DD[DIR_0PM *size_Mat];
- D.f[DIR_0PM ] = &DD[DIR_0MP *size_Mat];
- D.f[DIR_000] = &DD[DIR_000*size_Mat];
- D.f[DIR_PPP ] = &DD[DIR_MMM *size_Mat];
- D.f[DIR_MMP ] = &DD[DIR_PPM *size_Mat];
- D.f[DIR_PMP ] = &DD[DIR_MPM *size_Mat];
- D.f[DIR_MPP ] = &DD[DIR_PMM *size_Mat];
- D.f[DIR_PPM ] = &DD[DIR_MMP *size_Mat];
- D.f[DIR_MMM ] = &DD[DIR_PPP *size_Mat];
- D.f[DIR_PMM ] = &DD[DIR_MPP *size_Mat];
- D.f[DIR_MPM ] = &DD[DIR_PMP *size_Mat];
- }
- //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- real f1_E = (D.f[DIR_P00 ])[k1e ];
- real f1_W = (D.f[DIR_M00 ])[k1w ];
- real f1_N = (D.f[DIR_0P0 ])[k1n ];
- real f1_S = (D.f[DIR_0M0 ])[k1s ];
- real f1_T = (D.f[DIR_00P ])[k1t ];
- real f1_B = (D.f[DIR_00M ])[k1b ];
- real f1_NE = (D.f[DIR_PP0 ])[k1ne ];
- real f1_SW = (D.f[DIR_MM0 ])[k1sw ];
- real f1_SE = (D.f[DIR_PM0 ])[k1se ];
- real f1_NW = (D.f[DIR_MP0 ])[k1nw ];
- real f1_TE = (D.f[DIR_P0P ])[k1te ];
- real f1_BW = (D.f[DIR_M0M ])[k1bw ];
- real f1_BE = (D.f[DIR_P0M ])[k1be ];
- real f1_TW = (D.f[DIR_M0P ])[k1tw ];
- real f1_TN = (D.f[DIR_0PP ])[k1tn ];
- real f1_BS = (D.f[DIR_0MM ])[k1bs ];
- real f1_BN = (D.f[DIR_0PM ])[k1bn ];
- real f1_TS = (D.f[DIR_0MP ])[k1ts ];
- //real f1_ZERO = (D.f[DIR_000])[k1zero];
- real f1_TNE = (D.f[DIR_PPP ])[k1tne ];
- real f1_TSW = (D.f[DIR_MMP ])[k1tsw ];
- real f1_TSE = (D.f[DIR_PMP ])[k1tse ];
- real f1_TNW = (D.f[DIR_MPP ])[k1tnw ];
- real f1_BNE = (D.f[DIR_PPM ])[k1bne ];
- real f1_BSW = (D.f[DIR_MMM ])[k1bsw ];
- real f1_BSE = (D.f[DIR_PMM ])[k1bse ];
- real f1_BNW = (D.f[DIR_MPM ])[k1bnw ];
- //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- real f_E = (D.f[DIR_P00 ])[ke ];
- real f_W = (D.f[DIR_M00 ])[kw ];
- real f_N = (D.f[DIR_0P0 ])[kn ];
- real f_S = (D.f[DIR_0M0 ])[ks ];
- real f_T = (D.f[DIR_00P ])[kt ];
- real f_B = (D.f[DIR_00M ])[kb ];
- real f_NE = (D.f[DIR_PP0 ])[kne ];
- real f_SW = (D.f[DIR_MM0 ])[ksw ];
- real f_SE = (D.f[DIR_PM0 ])[kse ];
- real f_NW = (D.f[DIR_MP0 ])[knw ];
- real f_TE = (D.f[DIR_P0P ])[kte ];
- real f_BW = (D.f[DIR_M0M ])[kbw ];
- real f_BE = (D.f[DIR_P0M ])[kbe ];
- real f_TW = (D.f[DIR_M0P ])[ktw ];
- real f_TN = (D.f[DIR_0PP ])[ktn ];
- real f_BS = (D.f[DIR_0MM ])[kbs ];
- real f_BN = (D.f[DIR_0PM ])[kbn ];
- real f_TS = (D.f[DIR_0MP ])[kts ];
- //real f_ZERO = (D.f[DIR_000])[kzero];
- real f_TNE = (D.f[DIR_PPP ])[ktne ];
- real f_TSW = (D.f[DIR_MMP ])[ktsw ];
- real f_TSE = (D.f[DIR_PMP ])[ktse ];
- real f_TNW = (D.f[DIR_MPP ])[ktnw ];
- real f_BNE = (D.f[DIR_PPM ])[kbne ];
- real f_BSW = (D.f[DIR_MMM ])[kbsw ];
- real f_BSE = (D.f[DIR_PMM ])[kbse ];
- real f_BNW = (D.f[DIR_MPM ])[kbnw ];
- //////////////////////////////////////////////////////////////////////////
-
- //real vx1, vx2, vx3, drho;
- //real vx1, vx2, vx3, drho, drho1;
- //////////////////////////////////////////////////////////////////////////
- //Dichte
- // drho1 = f1_TSE + f1_TNW + f1_TNE + f1_TSW + f1_BSE + f1_BNW + f1_BNE + f1_BSW +
- // f1_BN + f1_TS + f1_TN + f1_BS + f1_BE + f1_TW + f1_TE + f1_BW + f1_SE + f1_NW + f1_NE + f1_SW +
- // f1_T + f1_B + f1_N + f1_S + f1_E + f1_W + ((D.f[DIR_000])[k1zero]);
- // drho = f_TSE + f_TNW + f_TNE + f_TSW + f_BSE + f_BNW + f_BNE + f_BSW +
- // f_BN + f_TS + f_TN + f_BS + f_BE + f_TW + f_TE + f_BW + f_SE + f_NW + f_NE + f_SW +
- // f_T + f_B + f_N + f_S + f_E + f_W + ((D.f[DIR_000])[kzero]);
-
- //////////////////////////////////////////////////////////////////////////
- //Ux
-
- //vx1 = (((f_TSE - f_BNW) - (f_TNW - f_BSE)) + ((f_TNE - f_BSW) - (f_TSW - f_BNE)) +
- // ((f_BE - f_TW) + (f_TE - f_BW)) + ((f_SE - f_NW) + (f_NE - f_SW)) +
- // (f_E - f_W)) /(one + drho);
-
-
- // vx2 = ((-(f_TSE - f_BNW) + (f_TNW - f_BSE)) + ((f_TNE - f_BSW) - (f_TSW - f_BNE)) +
- // ((f_BN - f_TS) + (f_TN - f_BS)) + (-(f_SE - f_NW) + (f_NE - f_SW)) +
- // (f_N - f_S)) /(one + drho);
-
- // vx3 = (((f_TSE - f_BNW) + (f_TNW - f_BSE)) + ((f_TNE - f_BSW) + (f_TSW - f_BNE)) +
- // (-(f_BN - f_TS) + (f_TN - f_BS)) + ((f_TE - f_BW) - (f_BE - f_TW)) +
- // (f_T - f_B)) /(one + drho);
+ if(k>=numberOfBCnodes) return;
+ ////////////////////////////////////////////////////////////////////////////////
+ //index
+ unsigned int KQK = k_Q[k];
+ // unsigned int kzero= KQK;
+ unsigned int ke = KQK;
+ unsigned int kw = neighborX[KQK];
+ unsigned int kn = KQK;
+ unsigned int ks = neighborY[KQK];
+ unsigned int kt = KQK;
+ unsigned int kb = neighborZ[KQK];
+ unsigned int ksw = neighborY[kw];
+ unsigned int kne = KQK;
+ unsigned int kse = ks;
+ unsigned int knw = kw;
+ unsigned int kbw = neighborZ[kw];
+ unsigned int kte = KQK;
+ unsigned int kbe = kb;
+ unsigned int ktw = kw;
+ unsigned int kbs = neighborZ[ks];
+ unsigned int ktn = KQK;
+ unsigned int kbn = kb;
+ unsigned int kts = ks;
+ unsigned int ktse = ks;
+ unsigned int kbnw = kbw;
+ unsigned int ktnw = kw;
+ unsigned int kbse = kbs;
+ unsigned int ktsw = ksw;
+ unsigned int kbne = kb;
+ unsigned int ktne = KQK;
+ unsigned int kbsw = neighborZ[ksw];
+ ////////////////////////////////////////////////////////////////////////////////
+ //index1
+ unsigned int K1QK = k_N[k];
+ //unsigned int k1zero= K1QK;
+ unsigned int k1e = K1QK;
+ unsigned int k1w = neighborX[K1QK];
+ unsigned int k1n = K1QK;
+ unsigned int k1s = neighborY[K1QK];
+ unsigned int k1t = K1QK;
+ unsigned int k1b = neighborZ[K1QK];
+ unsigned int k1sw = neighborY[k1w];
+ unsigned int k1ne = K1QK;
+ unsigned int k1se = k1s;
+ unsigned int k1nw = k1w;
+ unsigned int k1bw = neighborZ[k1w];
+ unsigned int k1te = K1QK;
+ unsigned int k1be = k1b;
+ unsigned int k1tw = k1w;
+ unsigned int k1bs = neighborZ[k1s];
+ unsigned int k1tn = K1QK;
+ unsigned int k1bn = k1b;
+ unsigned int k1ts = k1s;
+ unsigned int k1tse = k1s;
+ unsigned int k1bnw = k1bw;
+ unsigned int k1tnw = k1w;
+ unsigned int k1bse = k1bs;
+ unsigned int k1tsw = k1sw;
+ unsigned int k1bne = k1b;
+ unsigned int k1tne = K1QK;
+ unsigned int k1bsw = neighborZ[k1sw];
+ ////////////////////////////////////////////////////////////////////////////////
+ Distributions27 dist;
+ getPointersToDistributions(dist, distributions, numberOfLBnodes, isEvenTimestep);
+ real f[27], f1[27];
+ //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ f1[DIR_P00] = (dist.f[DIR_P00])[k1e ];
+ f1[DIR_M00] = (dist.f[DIR_M00])[k1w ];
+ f1[DIR_0P0] = (dist.f[DIR_0P0])[k1n ];
+ f1[DIR_0M0] = (dist.f[DIR_0M0])[k1s ];
+ f1[DIR_00P] = (dist.f[DIR_00P])[k1t ];
+ f1[DIR_00M] = (dist.f[DIR_00M])[k1b ];
+ f1[DIR_PP0] = (dist.f[DIR_PP0])[k1ne ];
+ f1[DIR_MM0] = (dist.f[DIR_MM0])[k1sw ];
+ f1[DIR_PM0] = (dist.f[DIR_PM0])[k1se ];
+ f1[DIR_MP0] = (dist.f[DIR_MP0])[k1nw ];
+ f1[DIR_P0P] = (dist.f[DIR_P0P])[k1te ];
+ f1[DIR_M0M] = (dist.f[DIR_M0M])[k1bw ];
+ f1[DIR_P0M] = (dist.f[DIR_P0M])[k1be ];
+ f1[DIR_M0P] = (dist.f[DIR_M0P])[k1tw ];
+ f1[DIR_0PP] = (dist.f[DIR_0PP])[k1tn ];
+ f1[DIR_0MM] = (dist.f[DIR_0MM])[k1bs ];
+ f1[DIR_0PM] = (dist.f[DIR_0PM])[k1bn ];
+ f1[DIR_0MP] = (dist.f[DIR_0MP])[k1ts ];
+ // f1[DIR_000] = (dist.f[DIR_000])[k1zero];
+ f1[DIR_PPP] = (dist.f[DIR_PPP])[k1tne ];
+ f1[DIR_MMP] = (dist.f[DIR_MMP])[k1tsw ];
+ f1[DIR_PMP] = (dist.f[DIR_PMP])[k1tse ];
+ f1[DIR_MPP] = (dist.f[DIR_MPP])[k1tnw ];
+ f1[DIR_PPM] = (dist.f[DIR_PPM])[k1bne ];
+ f1[DIR_MMM] = (dist.f[DIR_MMM])[k1bsw ];
+ f1[DIR_PMM] = (dist.f[DIR_PMM])[k1bse ];
+ f1[DIR_MPM] = (dist.f[DIR_MPM])[k1bnw ];
+ //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ f[DIR_P00] = (dist.f[DIR_P00])[ke ];
+ f[DIR_M00] = (dist.f[DIR_M00])[kw ];
+ f[DIR_0P0] = (dist.f[DIR_0P0])[kn ];
+ f[DIR_0M0] = (dist.f[DIR_0M0])[ks ];
+ f[DIR_00P] = (dist.f[DIR_00P])[kt ];
+ f[DIR_00M] = (dist.f[DIR_00M])[kb ];
+ f[DIR_PP0] = (dist.f[DIR_PP0])[kne ];
+ f[DIR_MM0] = (dist.f[DIR_MM0])[ksw ];
+ f[DIR_PM0] = (dist.f[DIR_PM0])[kse ];
+ f[DIR_MP0] = (dist.f[DIR_MP0])[knw ];
+ f[DIR_P0P] = (dist.f[DIR_P0P])[kte ];
+ f[DIR_M0M] = (dist.f[DIR_M0M])[kbw ];
+ f[DIR_P0M] = (dist.f[DIR_P0M])[kbe ];
+ f[DIR_M0P] = (dist.f[DIR_M0P])[ktw ];
+ f[DIR_0PP] = (dist.f[DIR_0PP])[ktn ];
+ f[DIR_0MM] = (dist.f[DIR_0MM])[kbs ];
+ f[DIR_0PM] = (dist.f[DIR_0PM])[kbn ];
+ f[DIR_0MP] = (dist.f[DIR_0MP])[kts ];
+ // f[DIR_000] = (dist.f[DIR_000])[kzero];
+ f[DIR_PPP] = (dist.f[DIR_PPP])[ktne ];
+ f[DIR_MMP] = (dist.f[DIR_MMP])[ktsw ];
+ f[DIR_PMP] = (dist.f[DIR_PMP])[ktse ];
+ f[DIR_MPP] = (dist.f[DIR_MPP])[ktnw ];
+ f[DIR_PPM] = (dist.f[DIR_PPM])[kbne ];
+ f[DIR_MMM] = (dist.f[DIR_MMM])[kbsw ];
+ f[DIR_PMM] = (dist.f[DIR_PMM])[kbse ];
+ f[DIR_MPM] = (dist.f[DIR_MPM])[kbnw ];
+ //////////////////////////////////////////////////////////////////////////
- //real cu_sq=c3o2*(vx1*vx1+vx2*vx2+vx3*vx3);
- // //////////////////////////////////////////////////////////////////////////
- ////real omega = om1;
- // real cusq = c3o2*(vx1*vx1+vx2*vx2+vx3*vx3);
- // //////////////////////////////////////////////////////////////////////////
- ////T�st MK
- ////if(vx1 < zero) vx1 = zero;
- // //////////////////////////////////////////////////////////////////////////
- // real fZERO = c8over27* (drho1-(one + drho1)*(cusq)) ;
- // real fE = c2over27* (drho1+(one + drho1)*(three*( vx1 )+c9over2*( vx1 )*( vx1 )-cusq));
- // real fW = c2over27* (drho1+(one + drho1)*(three*(-vx1 )+c9over2*(-vx1 )*(-vx1 )-cusq));
- // real fN = c2over27* (drho1+(one + drho1)*(three*( vx2 )+c9over2*( vx2 )*( vx2 )-cusq));
- // real fS = c2over27* (drho1+(one + drho1)*(three*( -vx2 )+c9over2*( -vx2 )*( -vx2 )-cusq));
- // real fT = c2over27* (drho1+(one + drho1)*(three*( vx3)+c9over2*( vx3)*( vx3)-cusq));
- // real fB = c2over27* (drho1+(one + drho1)*(three*( -vx3)+c9over2*( -vx3)*( -vx3)-cusq));
- // real fNE = c1over54* (drho1+(one + drho1)*(three*( vx1+vx2 )+c9over2*( vx1+vx2 )*( vx1+vx2 )-cusq));
- // real fSW = c1over54* (drho1+(one + drho1)*(three*(-vx1-vx2 )+c9over2*(-vx1-vx2 )*(-vx1-vx2 )-cusq));
- // real fSE = c1over54* (drho1+(one + drho1)*(three*( vx1-vx2 )+c9over2*( vx1-vx2 )*( vx1-vx2 )-cusq));
- // real fNW = c1over54* (drho1+(one + drho1)*(three*(-vx1+vx2 )+c9over2*(-vx1+vx2 )*(-vx1+vx2 )-cusq));
- // real fTE = c1over54* (drho1+(one + drho1)*(three*( vx1 +vx3)+c9over2*( vx1 +vx3)*( vx1 +vx3)-cusq));
- // real fBW = c1over54* (drho1+(one + drho1)*(three*(-vx1 -vx3)+c9over2*(-vx1 -vx3)*(-vx1 -vx3)-cusq));
- // real fBE = c1over54* (drho1+(one + drho1)*(three*( vx1 -vx3)+c9over2*( vx1 -vx3)*( vx1 -vx3)-cusq));
- // real fTW = c1over54* (drho1+(one + drho1)*(three*(-vx1 +vx3)+c9over2*(-vx1 +vx3)*(-vx1 +vx3)-cusq));
- // real fTN = c1over54* (drho1+(one + drho1)*(three*( vx2+vx3)+c9over2*( vx2+vx3)*( vx2+vx3)-cusq));
- // real fBS = c1over54* (drho1+(one + drho1)*(three*( -vx2-vx3)+c9over2*( -vx2-vx3)*( -vx2-vx3)-cusq));
- // real fBN = c1over54* (drho1+(one + drho1)*(three*( vx2-vx3)+c9over2*( vx2-vx3)*( vx2-vx3)-cusq));
- // real fTS = c1over54* (drho1+(one + drho1)*(three*( -vx2+vx3)+c9over2*( -vx2+vx3)*( -vx2+vx3)-cusq));
- // real fTNE = c1over216* (drho1+(one + drho1)*(three*( vx1+vx2+vx3)+c9over2*( vx1+vx2+vx3)*( vx1+vx2+vx3)-cusq));
- // real fBSW = c1over216* (drho1+(one + drho1)*(three*(-vx1-vx2-vx3)+c9over2*(-vx1-vx2-vx3)*(-vx1-vx2-vx3)-cusq));
- // real fBNE = c1over216* (drho1+(one + drho1)*(three*( vx1+vx2-vx3)+c9over2*( vx1+vx2-vx3)*( vx1+vx2-vx3)-cusq));
- // real fTSW = c1over216* (drho1+(one + drho1)*(three*(-vx1-vx2+vx3)+c9over2*(-vx1-vx2+vx3)*(-vx1-vx2+vx3)-cusq));
- // real fTSE = c1over216* (drho1+(one + drho1)*(three*( vx1-vx2+vx3)+c9over2*( vx1-vx2+vx3)*( vx1-vx2+vx3)-cusq));
- // real fBNW = c1over216* (drho1+(one + drho1)*(three*(-vx1+vx2-vx3)+c9over2*(-vx1+vx2-vx3)*(-vx1+vx2-vx3)-cusq));
- // real fBSE = c1over216* (drho1+(one + drho1)*(three*( vx1-vx2-vx3)+c9over2*( vx1-vx2-vx3)*( vx1-vx2-vx3)-cusq));
- // real fTNW = c1over216* (drho1+(one + drho1)*(three*(-vx1+vx2+vx3)+c9over2*(-vx1+vx2+vx3)*(-vx1+vx2+vx3)-cusq));
+ real cs = c1o1 / sqrtf(c3o1);
- real cs = c1o1 / sqrtf(c3o1);
- /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- //no velocity
- //////////////////////////////////////////
- f_E = f1_E * cs + (c1o1 - cs) * f_E ;
- f_W = f1_W * cs + (c1o1 - cs) * f_W ;
- f_N = f1_N * cs + (c1o1 - cs) * f_N ;
- f_S = f1_S * cs + (c1o1 - cs) * f_S ;
- f_T = f1_T * cs + (c1o1 - cs) * f_T ;
- f_B = f1_B * cs + (c1o1 - cs) * f_B ;
- f_NE = f1_NE * cs + (c1o1 - cs) * f_NE ;
- f_SW = f1_SW * cs + (c1o1 - cs) * f_SW ;
- f_SE = f1_SE * cs + (c1o1 - cs) * f_SE ;
- f_NW = f1_NW * cs + (c1o1 - cs) * f_NW ;
- f_TE = f1_TE * cs + (c1o1 - cs) * f_TE ;
- f_BW = f1_BW * cs + (c1o1 - cs) * f_BW ;
- f_BE = f1_BE * cs + (c1o1 - cs) * f_BE ;
- f_TW = f1_TW * cs + (c1o1 - cs) * f_TW ;
- f_TN = f1_TN * cs + (c1o1 - cs) * f_TN ;
- f_BS = f1_BS * cs + (c1o1 - cs) * f_BS ;
- f_BN = f1_BN * cs + (c1o1 - cs) * f_BN ;
- f_TS = f1_TS * cs + (c1o1 - cs) * f_TS ;
- f_TNE = f1_TNE * cs + (c1o1 - cs) * f_TNE ;
- f_TSW = f1_TSW * cs + (c1o1 - cs) * f_TSW ;
- f_TSE = f1_TSE * cs + (c1o1 - cs) * f_TSE ;
- f_TNW = f1_TNW * cs + (c1o1 - cs) * f_TNW ;
- f_BNE = f1_BNE * cs + (c1o1 - cs) * f_BNE ;
- f_BSW = f1_BSW * cs + (c1o1 - cs) * f_BSW ;
- f_BSE = f1_BSE * cs + (c1o1 - cs) * f_BSE ;
- f_BNW = f1_BNW * cs + (c1o1 - cs) * f_BNW ;
- /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-
- /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- //with velocity
- //if(true){//vx1 >= zero){
- // real csMvx = one / sqrtf(three) - vx1;
- // //real csMvy = one / sqrtf(three) - vx2;
- // ///////////////////////////////////////////
- // // X
- // f_W = f1_W * csMvx + (one - csMvx) * f_W ;//- c2over27 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // f_NW = f1_NW * csMvx + (one - csMvx) * f_NW ;//- c1over54 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // f_SW = f1_SW * csMvx + (one - csMvx) * f_SW ;//- c1over54 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // f_TW = f1_TW * csMvx + (one - csMvx) * f_TW ;//- c1over54 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // f_BW = f1_BW * csMvx + (one - csMvx) * f_BW ;//- c1over54 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // f_TNW = f1_TNW * csMvx + (one - csMvx) * f_TNW ;//- c1over216 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // f_TSW = f1_TSW * csMvx + (one - csMvx) * f_TSW ;//- c1over216 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // f_BNW = f1_BNW * csMvx + (one - csMvx) * f_BNW ;//- c1over216 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // f_BSW = f1_BSW * csMvx + (one - csMvx) * f_BSW ;//- c1over216 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx1);
- // ///////////////////////////////////////////
- // // Y
- // //f_S = f1_S * csMvy + (one - csMvy) * f_S ;//- c2over27 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_SE = f1_SE * csMvy + (one - csMvy) * f_SE ;//- c1over54 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_SW = f1_SW * csMvy + (one - csMvy) * f_SW ;//- c1over54 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_TS = f1_TS * csMvy + (one - csMvy) * f_TS ;//- c1over54 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_BS = f1_BS * csMvy + (one - csMvy) * f_BS ;//- c1over54 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_TSE = f1_TSE * csMvy + (one - csMvy) * f_TSE ;//- c1over216 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_TSW = f1_TSW * csMvy + (one - csMvy) * f_TSW ;//- c1over216 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_BSE = f1_BSE * csMvy + (one - csMvy) * f_BSE ;//- c1over216 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_BSW = f1_BSW * csMvy + (one - csMvy) * f_BSW ;//- c1over216 * ((drho + drho1)*c1o2-((drho + drho1)*c1o2 )*three*vx2);
- // //f_S = f1_S * csMvy + (one - csMvy) * f_S;
- // //f_SE = f1_SE * csMvy + (one - csMvy) * f_SE;
- // //f_SW = f1_SW * csMvy + (one - csMvy) * f_SW;
- // //f_TS = f1_TS * csMvy + (one - csMvy) * f_TS;
- // //f_BS = f1_BS * csMvy + (one - csMvy) * f_BS;
- // //f_TSE = f1_TSE * csMvy + (one - csMvy) * f_TSE;
- // //f_TSW = f1_TSW * csMvy + (one - csMvy) * f_TSW;
- // //f_BSE = f1_BSE * csMvy + (one - csMvy) * f_BSE;
- // //f_BSW = f1_BSW * csMvy + (one - csMvy) * f_BSW;
- // //////////////////////////////////////////////////////////////////////////
- //}
- //else
- //{
- // ///////////////////////////////////////////
- // // X
- // vx1 = vx1 * 0.9;
- // f_W = f_E - six * c2over27 * ( vx1 );
- // f_NW = f_SE - six * c1over54 * ( vx1-vx2 );
- // f_SW = f_NE - six * c1over54 * ( vx1+vx2 );
- // f_TW = f_BE - six * c1over54 * ( vx1 -vx3);
- // f_BW = f_TE - six * c1over54 * ( vx1 +vx3);
- // f_TNW = f_BSE - six * c1over216 * ( vx1-vx2-vx3);
- // f_TSW = f_BNE - six * c1over216 * ( vx1+vx2-vx3);
- // f_BNW = f_TSE - six * c1over216 * ( vx1-vx2+vx3);
- // f_BSW = f_TNE - six * c1over216 * ( vx1+vx2+vx3);
- // ///////////////////////////////////////////
- // // Y
- // //vx2 = vx2 * 0.9;
- // //f_S = f_N - six * c2over27 * ( vx2 );
- // //f_SE = f_NW - six * c1over54 * (-vx1+vx2 );
- // //f_SW = f_NE - six * c1over54 * ( vx1+vx2 );
- // //f_TS = f_BN - six * c1over54 * ( vx2-vx3);
- // //f_BS = f_TN - six * c1over54 * ( vx2+vx3);
- // //f_TSE = f_BNW - six * c1over216 * (-vx1+vx2-vx3);
- // //f_TSW = f_BNE - six * c1over216 * ( vx1+vx2-vx3);
- // //f_BSE = f_TNW - six * c1over216 * (-vx1+vx2+vx3);
- // //f_BSW = f_TNE - six * c1over216 * ( vx1+vx2+vx3);
- // ///////////////////////////////////////////
- //}
- /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-
- //////////////////////////////////////////////////////////////////////////
- if (isEvenTimestep==false)
- {
- D.f[DIR_P00 ] = &DD[DIR_P00 *size_Mat];
- D.f[DIR_M00 ] = &DD[DIR_M00 *size_Mat];
- D.f[DIR_0P0 ] = &DD[DIR_0P0 *size_Mat];
- D.f[DIR_0M0 ] = &DD[DIR_0M0 *size_Mat];
- D.f[DIR_00P ] = &DD[DIR_00P *size_Mat];
- D.f[DIR_00M ] = &DD[DIR_00M *size_Mat];
- D.f[DIR_PP0 ] = &DD[DIR_PP0 *size_Mat];
- D.f[DIR_MM0 ] = &DD[DIR_MM0 *size_Mat];
- D.f[DIR_PM0 ] = &DD[DIR_PM0 *size_Mat];
- D.f[DIR_MP0 ] = &DD[DIR_MP0 *size_Mat];
- D.f[DIR_P0P ] = &DD[DIR_P0P *size_Mat];
- D.f[DIR_M0M ] = &DD[DIR_M0M *size_Mat];
- D.f[DIR_P0M ] = &DD[DIR_P0M *size_Mat];
- D.f[DIR_M0P ] = &DD[DIR_M0P *size_Mat];
- D.f[DIR_0PP ] = &DD[DIR_0PP *size_Mat];
- D.f[DIR_0MM ] = &DD[DIR_0MM *size_Mat];
- D.f[DIR_0PM ] = &DD[DIR_0PM *size_Mat];
- D.f[DIR_0MP ] = &DD[DIR_0MP *size_Mat];
- D.f[DIR_000] = &DD[DIR_000*size_Mat];
- D.f[DIR_PPP ] = &DD[DIR_PPP *size_Mat];
- D.f[DIR_MMP ] = &DD[DIR_MMP *size_Mat];
- D.f[DIR_PMP ] = &DD[DIR_PMP *size_Mat];
- D.f[DIR_MPP ] = &DD[DIR_MPP *size_Mat];
- D.f[DIR_PPM ] = &DD[DIR_PPM *size_Mat];
- D.f[DIR_MMM ] = &DD[DIR_MMM *size_Mat];
- D.f[DIR_PMM ] = &DD[DIR_PMM *size_Mat];
- D.f[DIR_MPM ] = &DD[DIR_MPM *size_Mat];
- }
- else
- {
- D.f[DIR_M00 ] = &DD[DIR_P00 *size_Mat];
- D.f[DIR_P00 ] = &DD[DIR_M00 *size_Mat];
- D.f[DIR_0M0 ] = &DD[DIR_0P0 *size_Mat];
- D.f[DIR_0P0 ] = &DD[DIR_0M0 *size_Mat];
- D.f[DIR_00M ] = &DD[DIR_00P *size_Mat];
- D.f[DIR_00P ] = &DD[DIR_00M *size_Mat];
- D.f[DIR_MM0 ] = &DD[DIR_PP0 *size_Mat];
- D.f[DIR_PP0 ] = &DD[DIR_MM0 *size_Mat];
- D.f[DIR_MP0 ] = &DD[DIR_PM0 *size_Mat];
- D.f[DIR_PM0 ] = &DD[DIR_MP0 *size_Mat];
- D.f[DIR_M0M ] = &DD[DIR_P0P *size_Mat];
- D.f[DIR_P0P ] = &DD[DIR_M0M *size_Mat];
- D.f[DIR_M0P ] = &DD[DIR_P0M *size_Mat];
- D.f[DIR_P0M ] = &DD[DIR_M0P *size_Mat];
- D.f[DIR_0MM ] = &DD[DIR_0PP *size_Mat];
- D.f[DIR_0PP ] = &DD[DIR_0MM *size_Mat];
- D.f[DIR_0MP ] = &DD[DIR_0PM *size_Mat];
- D.f[DIR_0PM ] = &DD[DIR_0MP *size_Mat];
- D.f[DIR_000] = &DD[DIR_000*size_Mat];
- D.f[DIR_PPP ] = &DD[DIR_MMM *size_Mat];
- D.f[DIR_MMP ] = &DD[DIR_PPM *size_Mat];
- D.f[DIR_PMP ] = &DD[DIR_MPM *size_Mat];
- D.f[DIR_MPP ] = &DD[DIR_PMM *size_Mat];
- D.f[DIR_PPM ] = &DD[DIR_MMP *size_Mat];
- D.f[DIR_MMM ] = &DD[DIR_PPP *size_Mat];
- D.f[DIR_PMM ] = &DD[DIR_MPP *size_Mat];
- D.f[DIR_MPM ] = &DD[DIR_PMP *size_Mat];
- }
- //////////////////////////////////////////////////////////////////////////
- //__syncthreads();
- // -X
- //(D.f[DIR_P00 ])[ke ] = f_E ;
- //(D.f[DIR_PM0 ])[kse ] = f_SE ;
- //(D.f[DIR_PP0 ])[kne ] = f_NE ;
- //(D.f[DIR_P0M ])[kbe ] = f_BE ;
- //(D.f[DIR_P0P ])[kte ] = f_TE ;
- //(D.f[DIR_PMP ])[ktse ] = f_TSE ;
- //(D.f[DIR_PPP ])[ktne ] = f_TNE ;
- //(D.f[DIR_PMM ])[kbse ] = f_BSE ;
- //(D.f[DIR_PPM ])[kbne ] = f_BNE ;
- // X
- (D.f[DIR_M00 ])[kw ] = f_W ;
- (D.f[DIR_MM0 ])[ksw ] = f_SW ;
- (D.f[DIR_MP0 ])[knw ] = f_NW ;
- (D.f[DIR_M0M ])[kbw ] = f_BW ;
- (D.f[DIR_M0P ])[ktw ] = f_TW ;
- (D.f[DIR_MMP ])[ktsw ] = f_TSW ;
- (D.f[DIR_MPP ])[ktnw ] = f_TNW ;
- (D.f[DIR_MMM ])[kbsw ] = f_BSW ;
- (D.f[DIR_MPM ])[kbnw ] = f_BNW ;
- // Y
- //(D.f[DIR_0M0 ])[ks ] = f_S ;
- //(D.f[DIR_PM0 ])[kse ] = f_SE ;
- //(D.f[DIR_MM0 ])[ksw ] = f_SW ;
- //(D.f[DIR_0MP ])[kts ] = f_TS ;
- //(D.f[DIR_0MM ])[kbs ] = f_BS ;
- //(D.f[DIR_PMP ])[ktse ] = f_TSE ;
- //(D.f[DIR_MMP ])[ktsw ] = f_TSW ;
- //(D.f[DIR_PMM ])[kbse ] = f_BSE ;
- //(D.f[DIR_MMM ])[kbsw ] = f_BSW ;
- // Z
- //(D.f[DIR_00M ])[kb ] = f_B ;
- //(D.f[DIR_P0M ])[kbe ] = f_BE ;
- //(D.f[DIR_M0M ])[kbw ] = f_BW ;
- //(D.f[DIR_0PM ])[kbn ] = f_BN ;
- //(D.f[DIR_0MM ])[kbs ] = f_BS ;
- //(D.f[DIR_PPM ])[kbne ] = f_BNE ;
- //(D.f[DIR_MPM ])[kbnw ] = f_BNW ;
- //(D.f[DIR_PMM ])[kbse ] = f_BSE ;
- //(D.f[DIR_MMM ])[kbsw ] = f_BSW ;
- //////////////////////////////////////////////////////////////////////////
+ //////////////////////////////////////////////////////////////////////////
+ getPointersToDistributions(dist, distributions, numberOfLBnodes, !isEvenTimestep);
+ switch(direction)
+ {
+ case MZZ:
+ (dist.f[DIR_P00])[ke ] = computeOutflowDistribution(f, f1, DIR_P00, cs);
+ (dist.f[DIR_PM0])[kse ] = computeOutflowDistribution(f, f1, DIR_PM0, cs);
+ (dist.f[DIR_PP0])[kne ] = computeOutflowDistribution(f, f1, DIR_PP0, cs);
+ (dist.f[DIR_P0M])[kbe ] = computeOutflowDistribution(f, f1, DIR_P0M, cs);
+ (dist.f[DIR_P0P])[kte ] = computeOutflowDistribution(f, f1, DIR_P0P, cs);
+ (dist.f[DIR_PMP])[ktse ] = computeOutflowDistribution(f, f1, DIR_PMP, cs);
+ (dist.f[DIR_PPP])[ktne ] = computeOutflowDistribution(f, f1, DIR_PPP, cs);
+ (dist.f[DIR_PMM])[kbse ] = computeOutflowDistribution(f, f1, DIR_PMM, cs);
+ (dist.f[DIR_PPM])[kbne ] = computeOutflowDistribution(f, f1, DIR_PPM, cs);
+ break;
+
+ case PZZ:
+ (dist.f[DIR_M00])[kw ] = computeOutflowDistribution(f, f1, DIR_M00, cs);
+ (dist.f[DIR_MM0])[ksw ] = computeOutflowDistribution(f, f1, DIR_MM0, cs);
+ (dist.f[DIR_MP0])[knw ] = computeOutflowDistribution(f, f1, DIR_MP0, cs);
+ (dist.f[DIR_M0M])[kbw ] = computeOutflowDistribution(f, f1, DIR_M0M, cs);
+ (dist.f[DIR_M0P])[ktw ] = computeOutflowDistribution(f, f1, DIR_M0P, cs);
+ (dist.f[DIR_MMP])[ktsw ] = computeOutflowDistribution(f, f1, DIR_MMP, cs);
+ (dist.f[DIR_MPP])[ktnw ] = computeOutflowDistribution(f, f1, DIR_MPP, cs);
+ (dist.f[DIR_MMM])[kbsw ] = computeOutflowDistribution(f, f1, DIR_MMM, cs);
+ (dist.f[DIR_MPM])[kbnw ] = computeOutflowDistribution(f, f1, DIR_MPM, cs);
+ break;
+
+ case ZMZ:
+ (dist.f[DIR_0P0])[kn ] = computeOutflowDistribution(f, f1, DIR_0P0, cs);
+ (dist.f[DIR_PP0])[kne ] = computeOutflowDistribution(f, f1, DIR_PP0, cs);
+ (dist.f[DIR_MP0])[knw ] = computeOutflowDistribution(f, f1, DIR_MP0, cs);
+ (dist.f[DIR_0PP])[ktn ] = computeOutflowDistribution(f, f1, DIR_0PP, cs);
+ (dist.f[DIR_0PM])[kbn ] = computeOutflowDistribution(f, f1, DIR_0PM, cs);
+ (dist.f[DIR_PPP])[ktne ] = computeOutflowDistribution(f, f1, DIR_PPP, cs);
+ (dist.f[DIR_MPP])[ktnw ] = computeOutflowDistribution(f, f1, DIR_MPP, cs);
+ (dist.f[DIR_PPM])[kbne ] = computeOutflowDistribution(f, f1, DIR_PPM, cs);
+ (dist.f[DIR_MPM])[kbnw ] = computeOutflowDistribution(f, f1, DIR_MPM, cs);
+ break;
+
+ case ZPZ:
+ (dist.f[DIR_0M0])[ks ] = computeOutflowDistribution(f, f1, DIR_0M0, cs);
+ (dist.f[DIR_PM0])[kse ] = computeOutflowDistribution(f, f1, DIR_PM0, cs);
+ (dist.f[DIR_MM0])[ksw ] = computeOutflowDistribution(f, f1, DIR_MM0, cs);
+ (dist.f[DIR_0MP])[kts ] = computeOutflowDistribution(f, f1, DIR_0MP, cs);
+ (dist.f[DIR_0MM])[kbs ] = computeOutflowDistribution(f, f1, DIR_0MM, cs);
+ (dist.f[DIR_PMP])[ktse ] = computeOutflowDistribution(f, f1, DIR_PMP, cs);
+ (dist.f[DIR_MMP])[ktsw ] = computeOutflowDistribution(f, f1, DIR_MMP, cs);
+ (dist.f[DIR_PMM])[kbse ] = computeOutflowDistribution(f, f1, DIR_PMM, cs);
+ (dist.f[DIR_MMM])[kbsw ] = computeOutflowDistribution(f, f1, DIR_MMM, cs);
+ break;
+
+ case ZZM:
+ (dist.f[DIR_00P])[kt ] = computeOutflowDistribution(f, f1, DIR_00P, cs);
+ (dist.f[DIR_P0P])[kte ] = computeOutflowDistribution(f, f1, DIR_P0P, cs);
+ (dist.f[DIR_M0P])[ktw ] = computeOutflowDistribution(f, f1, DIR_M0P, cs);
+ (dist.f[DIR_0PP])[ktn ] = computeOutflowDistribution(f, f1, DIR_0PP, cs);
+ (dist.f[DIR_0MP])[kts ] = computeOutflowDistribution(f, f1, DIR_0MP, cs);
+ (dist.f[DIR_PPP])[ktne ] = computeOutflowDistribution(f, f1, DIR_PPP, cs);
+ (dist.f[DIR_MPP])[ktnw ] = computeOutflowDistribution(f, f1, DIR_MPP, cs);
+ (dist.f[DIR_PMP])[ktse ] = computeOutflowDistribution(f, f1, DIR_PMP, cs);
+ (dist.f[DIR_MMP])[ktsw ] = computeOutflowDistribution(f, f1, DIR_MMP, cs);
+ break;
+
+ case ZZP:
+ (dist.f[DIR_00M])[kb ] = computeOutflowDistribution(f, f1, DIR_00M, cs);
+ (dist.f[DIR_P0M])[kbe ] = computeOutflowDistribution(f, f1, DIR_P0M, cs);
+ (dist.f[DIR_M0M])[kbw ] = computeOutflowDistribution(f, f1, DIR_M0M, cs);
+ (dist.f[DIR_0PM])[kbn ] = computeOutflowDistribution(f, f1, DIR_0PM, cs);
+ (dist.f[DIR_0MM])[kbs ] = computeOutflowDistribution(f, f1, DIR_0MM, cs);
+ (dist.f[DIR_PPM])[kbne ] = computeOutflowDistribution(f, f1, DIR_PPM, cs);
+ (dist.f[DIR_MPM])[kbnw ] = computeOutflowDistribution(f, f1, DIR_MPM, cs);
+ (dist.f[DIR_PMM])[kbse ] = computeOutflowDistribution(f, f1, DIR_PMM, cs);
+ (dist.f[DIR_MMM])[kbsw ] = computeOutflowDistribution(f, f1, DIR_MMM, cs);
+ break;
+ default:
+ break;
}
}
-////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+__host__ __device__ real computeOutflowDistribution(const real* const &f, const real* const &f1, const int dir, const real rhoCorrection, const real cs, const real weight)
+{
+ return f1[dir ] * cs + (c1o1 - cs) * f[dir ] - weight *rhoCorrection;
+}
+__global__ void QPressZeroRhoOutflowDevice27( real* rhoBC,
+ real* distributions,
+ int* k_Q,
+ int* k_N,
+ int numberOfBCnodes,
+ real om1,
+ unsigned int* neighborX,
+ unsigned int* neighborY,
+ unsigned int* neighborZ,
+ unsigned int numberOfLBnodes,
+ bool isEvenTimestep,
+ int direction,
+ real densityCorrectionFactor)
+{
+ ////////////////////////////////////////////////////////////////////////////////
+ const unsigned k = vf::gpu::getNodeIndex();
+
+ //////////////////////////////////////////////////////////////////////////
+ if(k>=numberOfBCnodes) return;
+ ////////////////////////////////////////////////////////////////////////////////
+ //index
+ uint k_000 = k_Q[k];
+ uint k_M00 = neighborX[k_000];
+ uint k_0M0 = neighborY[k_000];
+ uint k_00M = neighborZ[k_000];
+ uint k_MM0 = neighborY[k_M00];
+ uint k_M0M = neighborZ[k_M00];
+ uint k_0MM = neighborZ[k_0M0];
+ uint k_MMM = neighborZ[k_MM0];
+ ////////////////////////////////////////////////////////////////////////////////
+ //index of neighbor
+ uint kN_000 = k_N[k];
+ uint kN_M00 = neighborX[k_000];
+ uint kN_0M0 = neighborY[k_000];
+ uint kN_00M = neighborZ[k_000];
+ uint kN_MM0 = neighborY[k_M00];
+ uint kN_M0M = neighborZ[k_M00];
+ uint kN_0MM = neighborZ[k_0M0];
+ uint kN_MMM = neighborZ[k_MM0];
+ ////////////////////////////////////////////////////////////////////////////////
+ Distributions27 dist;
+ getPointersToDistributions(dist, distributions, numberOfLBnodes, isEvenTimestep);
+ real f[27], fN[27];
+ //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ f[DIR_000] = (dist.f[DIR_000])[k_000];
+ f[DIR_P00] = (dist.f[DIR_P00])[k_000];
+ f[DIR_M00] = (dist.f[DIR_M00])[k_M00];
+ f[DIR_0P0] = (dist.f[DIR_0P0])[k_000];
+ f[DIR_0M0] = (dist.f[DIR_0M0])[k_0M0];
+ f[DIR_00P] = (dist.f[DIR_00P])[k_000];
+ f[DIR_00M] = (dist.f[DIR_00M])[k_00M];
+ f[DIR_PP0] = (dist.f[DIR_PP0])[k_000];
+ f[DIR_MM0] = (dist.f[DIR_MM0])[k_MM0];
+ f[DIR_PM0] = (dist.f[DIR_PM0])[k_0M0];
+ f[DIR_MP0] = (dist.f[DIR_MP0])[k_M00];
+ f[DIR_P0P] = (dist.f[DIR_P0P])[k_000];
+ f[DIR_M0M] = (dist.f[DIR_M0M])[k_M0M];
+ f[DIR_P0M] = (dist.f[DIR_P0M])[k_00M];
+ f[DIR_M0P] = (dist.f[DIR_M0P])[k_M00];
+ f[DIR_0PP] = (dist.f[DIR_0PP])[k_000];
+ f[DIR_0MM] = (dist.f[DIR_0MM])[k_0MM];
+ f[DIR_0PM] = (dist.f[DIR_0PM])[k_00M];
+ f[DIR_0MP] = (dist.f[DIR_0MP])[k_0M0];
+ f[DIR_PPP] = (dist.f[DIR_PPP])[k_000];
+ f[DIR_MPP] = (dist.f[DIR_MPP])[k_M00];
+ f[DIR_PMP] = (dist.f[DIR_PMP])[k_0M0];
+ f[DIR_MMP] = (dist.f[DIR_MMP])[k_MM0];
+ f[DIR_PPM] = (dist.f[DIR_PPM])[k_00M];
+ f[DIR_MPM] = (dist.f[DIR_MPM])[k_M0M];
+ f[DIR_PMM] = (dist.f[DIR_PMM])[k_0MM];
+ f[DIR_MMM] = (dist.f[DIR_MMM])[k_MMM];
+ //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+ fN[DIR_000] = (dist.f[DIR_000])[kN_000];
+ fN[DIR_P00] = (dist.f[DIR_P00])[kN_000];
+ fN[DIR_M00] = (dist.f[DIR_M00])[kN_M00];
+ fN[DIR_0P0] = (dist.f[DIR_0P0])[kN_000];
+ fN[DIR_0M0] = (dist.f[DIR_0M0])[kN_0M0];
+ fN[DIR_00P] = (dist.f[DIR_00P])[kN_000];
+ fN[DIR_00M] = (dist.f[DIR_00M])[kN_00M];
+ fN[DIR_PP0] = (dist.f[DIR_PP0])[kN_000];
+ fN[DIR_MM0] = (dist.f[DIR_MM0])[kN_MM0];
+ fN[DIR_PM0] = (dist.f[DIR_PM0])[kN_0M0];
+ fN[DIR_MP0] = (dist.f[DIR_MP0])[kN_M00];
+ fN[DIR_P0P] = (dist.f[DIR_P0P])[kN_000];
+ fN[DIR_M0M] = (dist.f[DIR_M0M])[kN_M0M];
+ fN[DIR_P0M] = (dist.f[DIR_P0M])[kN_00M];
+ fN[DIR_M0P] = (dist.f[DIR_M0P])[kN_M00];
+ fN[DIR_0PP] = (dist.f[DIR_0PP])[kN_000];
+ fN[DIR_0MM] = (dist.f[DIR_0MM])[kN_0MM];
+ fN[DIR_0PM] = (dist.f[DIR_0PM])[kN_00M];
+ fN[DIR_0MP] = (dist.f[DIR_0MP])[kN_0M0];
+ fN[DIR_PPP] = (dist.f[DIR_PPP])[kN_000];
+ fN[DIR_MPP] = (dist.f[DIR_MPP])[kN_M00];
+ fN[DIR_PMP] = (dist.f[DIR_PMP])[kN_0M0];
+ fN[DIR_MMP] = (dist.f[DIR_MMP])[kN_MM0];
+ fN[DIR_PPM] = (dist.f[DIR_PPM])[kN_00M];
+ fN[DIR_MPM] = (dist.f[DIR_MPM])[kN_M0M];
+ fN[DIR_PMM] = (dist.f[DIR_PMM])[kN_0MM];
+ fN[DIR_MMM] = (dist.f[DIR_MMM])[kN_MMM];
+ //////////////////////////////////////////////////////////////////////////
+ real drho = vf::lbm::getDensity(f);
+
+ real rhoCorrection = densityCorrectionFactor*drho;
+
+ real cs = c1o1 / sqrtf(c3o1);
+ getPointersToDistributions(dist, distributions, numberOfLBnodes, !isEvenTimestep);
+ switch(direction)
+ {
+ case MZZ:
+ (dist.f[DIR_P00])[k_000] = computeOutflowDistribution(f, fN, DIR_P00 , rhoCorrection, cs, c2o27);
+ (dist.f[DIR_PM0])[k_0M0] = computeOutflowDistribution(f, fN, DIR_PM0, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_PP0])[k_000] = computeOutflowDistribution(f, fN, DIR_PP0, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_P0M])[k_00M] = computeOutflowDistribution(f, fN, DIR_P0M, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_P0P])[k_000] = computeOutflowDistribution(f, fN, DIR_P0P, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_PMP])[k_0M0] = computeOutflowDistribution(f, fN, DIR_PMP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_PPP])[k_000] = computeOutflowDistribution(f, fN, DIR_PPP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_PMM])[k_0MM] = computeOutflowDistribution(f, fN, DIR_PMM, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_PPM])[k_00M] = computeOutflowDistribution(f, fN, DIR_PPM, rhoCorrection, cs, c1o216);
+ break;
+
+ case PZZ:
+ (dist.f[DIR_M00])[k_M00] = computeOutflowDistribution(f, fN, DIR_M00, rhoCorrection, cs, c2o27);
+ (dist.f[DIR_MM0])[k_MM0] = computeOutflowDistribution(f, fN, DIR_MM0, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_MP0])[k_M00] = computeOutflowDistribution(f, fN, DIR_MP0, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_M0M])[k_M0M] = computeOutflowDistribution(f, fN, DIR_M0M, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_M0P])[k_M00] = computeOutflowDistribution(f, fN, DIR_M0P, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_MMP])[k_MM0] = computeOutflowDistribution(f, fN, DIR_MMP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MPP])[k_M00] = computeOutflowDistribution(f, fN, DIR_MPP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MMM])[k_MMM] = computeOutflowDistribution(f, fN, DIR_MMM, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MPM])[k_M0M] = computeOutflowDistribution(f, fN, DIR_MPM, rhoCorrection, cs, c1o216);
+ break;
+
+ case ZMZ:
+ (dist.f[DIR_0P0])[k_000] = computeOutflowDistribution(f, fN, DIR_0P0, rhoCorrection, cs, c2o27);
+ (dist.f[DIR_PP0])[k_000] = computeOutflowDistribution(f, fN, DIR_PP0, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_MP0])[k_M00] = computeOutflowDistribution(f, fN, DIR_MP0, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_0PP])[k_000] = computeOutflowDistribution(f, fN, DIR_0PP, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_0PM])[k_00M] = computeOutflowDistribution(f, fN, DIR_0PM, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_PPP])[k_000] = computeOutflowDistribution(f, fN, DIR_PPP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MPP])[k_M00] = computeOutflowDistribution(f, fN, DIR_MPP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_PPM])[k_00M] = computeOutflowDistribution(f, fN, DIR_PPM, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MPM])[k_M0M] = computeOutflowDistribution(f, fN, DIR_MPM, rhoCorrection, cs, c1o216);
+ break;
+
+ case ZPZ:
+ (dist.f[DIR_0M0])[k_0M0] =computeOutflowDistribution(f, fN, DIR_0M0, rhoCorrection, cs, c2o27);
+ (dist.f[DIR_PM0])[k_0M0] =computeOutflowDistribution(f, fN, DIR_PM0, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_MM0])[k_MM0] =computeOutflowDistribution(f, fN, DIR_MM0, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_0MP])[k_0M0] =computeOutflowDistribution(f, fN, DIR_0MP, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_0MM])[k_0MM] =computeOutflowDistribution(f, fN, DIR_0MM, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_PMP])[k_0M0] =computeOutflowDistribution(f, fN, DIR_PMP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MMP])[k_MM0] =computeOutflowDistribution(f, fN, DIR_MMP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_PMM])[k_0MM] =computeOutflowDistribution(f, fN, DIR_PMM, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MMM])[k_MMM] =computeOutflowDistribution(f, fN, DIR_MMM, rhoCorrection, cs, c1o216);
+ break;
+
+ case ZZM:
+ (dist.f[DIR_00P])[k_000] = computeOutflowDistribution(f, fN, DIR_00P, rhoCorrection, cs, c2o27);
+ (dist.f[DIR_P0P])[k_000] = computeOutflowDistribution(f, fN, DIR_P0P, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_M0P])[k_M00] = computeOutflowDistribution(f, fN, DIR_M0P, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_0PP])[k_000] = computeOutflowDistribution(f, fN, DIR_0PP, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_0MP])[k_0M0] = computeOutflowDistribution(f, fN, DIR_0MP, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_PPP])[k_000] = computeOutflowDistribution(f, fN, DIR_PPP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MPP])[k_M00] = computeOutflowDistribution(f, fN, DIR_MPP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_PMP])[k_0M0] = computeOutflowDistribution(f, fN, DIR_PMP, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MMP])[k_MM0] = computeOutflowDistribution(f, fN, DIR_MMP, rhoCorrection, cs, c1o216);
+ break;
+
+ case ZZP:
+ (dist.f[DIR_00M])[k_00M] = computeOutflowDistribution(f, fN, DIR_00M, rhoCorrection, cs, c2o27);
+ (dist.f[DIR_P0M])[k_00M] = computeOutflowDistribution(f, fN, DIR_P0M, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_M0M])[k_M0M] = computeOutflowDistribution(f, fN, DIR_M0M, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_0PM])[k_00M] = computeOutflowDistribution(f, fN, DIR_0PM, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_0MM])[k_0MM] = computeOutflowDistribution(f, fN, DIR_0MM, rhoCorrection, cs, c1o54);
+ (dist.f[DIR_PPM])[k_00M] = computeOutflowDistribution(f, fN, DIR_PPM, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MPM])[k_M0M] = computeOutflowDistribution(f, fN, DIR_MPM, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_PMM])[k_0MM] = computeOutflowDistribution(f, fN, DIR_PMM, rhoCorrection, cs, c1o216);
+ (dist.f[DIR_MMM])[k_MMM] = computeOutflowDistribution(f, fN, DIR_MMM, rhoCorrection, cs, c1o216);
+ break;
+ default:
+ break;
+ }
+}
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
diff --git a/src/gpu/VirtualFluids_GPU/GPU/TurbulentViscosityInlines.cuh b/src/gpu/VirtualFluids_GPU/GPU/TurbulentViscosityInlines.cuh
index eb301515527a9e8a3056676b0d4dffe8197c7dbe..58856f624fa1dfd2488c3061721e9dac53a67d07 100644
--- a/src/gpu/VirtualFluids_GPU/GPU/TurbulentViscosityInlines.cuh
+++ b/src/gpu/VirtualFluids_GPU/GPU/TurbulentViscosityInlines.cuh
@@ -53,8 +53,8 @@ __inline__ __device__ real calcTurbulentViscosityQR(real C, real dxux, real dyuy
//! Second invariant of the strain-rate tensor
real Q = c1o2*( dxux*dxux + dyuy*dyuy + dzuz*dzuz ) + c1o4*( Dxy*Dxy + Dxz*Dxz + Dyz*Dyz);
//! Third invariant of the strain-rate tensor (determinant)
- real R = - dxux*dyuy*dzuz - c1o4*( Dxy*Dxz*Dyz + dxux*Dyz*Dyz + dyuy*Dxz*Dxz + dzuz*Dxy*Dxy );
-
+ // real R = - dxux*dyuy*dzuz - c1o4*( Dxy*Dxz*Dyz + dxux*Dyz*Dyz + dyuy*Dxz*Dxz + dzuz*Dxy*Dxy );
+ real R = - dxux*dyuy*dzuz + c1o4*( -Dxy*Dxz*Dyz + dxux*Dyz*Dyz + dyuy*Dxz*Dxz + dzuz*Dxy*Dxy );
return C * max(R, c0o1) / Q;
}
diff --git a/src/gpu/VirtualFluids_GPU/GPU/TurbulentViscosityKernels.cu b/src/gpu/VirtualFluids_GPU/GPU/TurbulentViscosityKernels.cu
index 3719ca3712e6f63a77f62bf314af7d19eea01f4c..a8f02fee717caf7f67624243b873fe993b5c7927 100644
--- a/src/gpu/VirtualFluids_GPU/GPU/TurbulentViscosityKernels.cu
+++ b/src/gpu/VirtualFluids_GPU/GPU/TurbulentViscosityKernels.cu
@@ -38,6 +38,7 @@
#include <cuda_runtime.h>
#include <helper_cuda.h>
#include "LBM/LB.h"
+#include "Kernel/Utilities/DistributionHelper.cuh"
using namespace vf::lbm::constant;
@@ -64,15 +65,7 @@ __global__ void calcAMD(real* vx,
uint size_Mat,
real SGSConstant)
{
-
- const uint x = threadIdx.x;
- const uint y = blockIdx.x;
- const uint z = blockIdx.y;
-
- const uint nx = blockDim.x;
- const uint ny = gridDim.x;
-
- const uint k = nx*(ny*z + y) + x;
+ const uint k = vf::gpu::getNodeIndex();
if(k >= size_Mat) return;
if(typeOfGridNode[k] != GEO_FLUID) return;
@@ -102,7 +95,7 @@ __global__ void calcAMD(real* vx,
(dvxdx*dvzdx + dvxdy*dvzdy + dvxdz*dvzdz) * (dvxdz+dvzdx) +
(dvydx*dvzdx + dvydy*dvzdy + dvydz*dvzdz) * (dvydz+dvzdy);
- turbulentViscosity[k] = max(c0o1,-SGSConstant*enumerator)/denominator;
+ turbulentViscosity[k] = denominator != c0o1 ? max(c0o1,-SGSConstant*enumerator)/denominator : c0o1;
}
void calcTurbulentViscosityAMD(Parameter* para, int level)
diff --git a/src/gpu/VirtualFluids_GPU/Init/PositionReader.cpp b/src/gpu/VirtualFluids_GPU/Init/PositionReader.cpp
index 4e5a862d3fd1ed19109073aae0fe4c731f7f3e91..04052b38cbfb59bfe620354b62da70402d74bdbd 100644
--- a/src/gpu/VirtualFluids_GPU/Init/PositionReader.cpp
+++ b/src/gpu/VirtualFluids_GPU/Init/PositionReader.cpp
@@ -3,6 +3,7 @@
#include "Parameter/Parameter.h"
#include <basics/utilities/UbFileInputASCII.h>
+
using namespace vf::lbm::dir;
//////////////////////////////////////////////////////////////////////////
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernel.h b/src/gpu/VirtualFluids_GPU/Kernel/Kernel.h
index 9f9f7539bc5a1e28612d956ca32234c5a3589f8a..50b4460d774010ea7d7b98cfa6fa505cdfeb88c2 100644
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernel.h
+++ b/src/gpu/VirtualFluids_GPU/Kernel/Kernel.h
@@ -3,8 +3,11 @@
#include <vector>
+#include "LBM/LB.h"
+
#include "Kernel/Utilities/KernelGroup.h"
#include "PreProcessor/PreProcessorType.h"
+#include "Parameter/CudaStreamManager.h"
#include <helper_cuda.h>
@@ -13,7 +16,7 @@ class Kernel
public:
virtual ~Kernel() = default;
virtual void run() = 0;
- virtual void runOnIndices(const unsigned int *indices, unsigned int size_indices, int stream = -1) = 0; //if stream == -1: run on default stream
+ virtual void runOnIndices(const unsigned int *indices, unsigned int size_indices, CollisionTemplate collisionTemplate, CudaStreamIndex streamIdx=CudaStreamIndex::Legacy) = 0;
virtual bool checkParameter() = 0;
virtual std::vector<PreProcessorType> getPreProcessorTypes() = 0;
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/KernelImp.cpp b/src/gpu/VirtualFluids_GPU/Kernel/KernelImp.cpp
index 630aaf7339afc2907ab6bfbf65bd5fc55f75e215..9bd3945aa81147d03be2b1eac3ddec7c24d71532 100644
--- a/src/gpu/VirtualFluids_GPU/Kernel/KernelImp.cpp
+++ b/src/gpu/VirtualFluids_GPU/Kernel/KernelImp.cpp
@@ -1,9 +1,11 @@
#include "KernelImp.h"
+#include "LBM/LB.h"
+
#include "Kernel/Utilities/CheckParameterStrategy/CheckParameterStrategy.h"
-void KernelImp::runOnIndices(const unsigned int *indices, unsigned int size_indices, int stream)
+void KernelImp::runOnIndices(const unsigned int *indices, unsigned int size_indices, CollisionTemplate collisionTemplate, CudaStreamIndex streamIndex)
{
printf("Method not implemented for this Kernel \n");
}
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/KernelImp.h b/src/gpu/VirtualFluids_GPU/Kernel/KernelImp.h
index 0141ddda7e9579cc84148d26727ed81c084ea0c5..a96c2c123472ca33f635273e06a5bf36a745654d 100644
--- a/src/gpu/VirtualFluids_GPU/Kernel/KernelImp.h
+++ b/src/gpu/VirtualFluids_GPU/Kernel/KernelImp.h
@@ -1,6 +1,8 @@
#ifndef KERNEL_IMP_H
#define KERNEL_IMP_H
+#include "LBM/LB.h"
+
#include "Kernel.h"
#include <memory>
@@ -9,12 +11,12 @@
class CheckParameterStrategy;
class Parameter;
-
+class CudaStreamManager;
class KernelImp : public Kernel
{
public:
virtual void run() = 0;
- virtual void runOnIndices(const unsigned int *indices, unsigned int size_indices, int stream = -1);
+ virtual void runOnIndices(const unsigned int *indices, unsigned int size_indices, CollisionTemplate collisionTemplate, CudaStreamIndex streamIndex=CudaStreamIndex::Legacy);
bool checkParameter();
std::vector<PreProcessorType> getPreProcessorTypes();
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17.cu
new file mode 100644
index 0000000000000000000000000000000000000000..70b0c4352afee850a4e17243979268bd126b7b4a
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17.cu
@@ -0,0 +1,140 @@
+#include "CumulantK17.h"
+#include <logger/Logger.h>
+#include "Parameter/Parameter.h"
+#include "Parameter/CudaStreamManager.h"
+#include "CumulantK17_Device.cuh"
+
+#include <cuda.h>
+
+template<TurbulenceModel turbulenceModel>
+std::shared_ptr< CumulantK17<turbulenceModel> > CumulantK17<turbulenceModel>::getNewInstance(std::shared_ptr<Parameter> para, int level)
+{
+ return std::shared_ptr<CumulantK17<turbulenceModel> >(new CumulantK17<turbulenceModel>(para,level));
+}
+
+template<TurbulenceModel turbulenceModel>
+void CumulantK17<turbulenceModel>::run()
+{
+ LB_Kernel_CumulantK17 < turbulenceModel, false, false > <<< cudaGrid.grid, cudaGrid.threads >>>( para->getParD(level)->omega,
+ para->getParD(level)->neighborX, para->getParD(level)->neighborY, para->getParD(level)->neighborZ,
+ para->getParD(level)->distributions.f[0],
+ para->getParD(level)->rho,
+ para->getParD(level)->velocityX, para->getParD(level)->velocityY, para->getParD(level)->velocityZ,
+ para->getParD(level)->turbViscosity,
+ para->getSGSConstant(),
+ (unsigned long)para->getParD(level)->numberOfNodes,
+ level,
+ para->getForcesDev(),
+ para->getParD(level)->forceX_SP, para->getParD(level)->forceY_SP, para->getParD(level)->forceZ_SP,
+ para->getQuadricLimitersDev(),
+ para->getParD(level)->isEvenTimestep,
+ para->getParD(level)->taggedFluidNodeIndices[CollisionTemplate::Default],
+ para->getParD(level)->numberOfTaggedFluidNodes[CollisionTemplate::Default]);
+
+ getLastCudaError("LB_Kernel_CumulantK17 execution failed");
+}
+
+template<TurbulenceModel turbulenceModel>
+void CumulantK17<turbulenceModel>::runOnIndices( const unsigned int *indices, unsigned int size_indices, CollisionTemplate collisionTemplate, CudaStreamIndex streamIndex )
+{
+ cudaStream_t stream = para->getStreamManager()->getStream(streamIndex);
+
+ switch (collisionTemplate)
+ {
+ case CollisionTemplate::Default:
+ LB_Kernel_CumulantK17 < turbulenceModel, false, false > <<< cudaGrid.grid, cudaGrid.threads, 0, stream >>>(para->getParD(level)->omega,
+ para->getParD(level)->neighborX, para->getParD(level)->neighborY, para->getParD(level)->neighborZ,
+ para->getParD(level)->distributions.f[0],
+ para->getParD(level)->rho,
+ para->getParD(level)->velocityX, para->getParD(level)->velocityY, para->getParD(level)->velocityZ,
+ para->getParD(level)->turbViscosity,
+ para->getSGSConstant(),
+ (unsigned long)para->getParD(level)->numberOfNodes,
+ level,
+ para->getForcesDev(),
+ para->getParD(level)->forceX_SP, para->getParD(level)->forceY_SP, para->getParD(level)->forceZ_SP,
+ para->getQuadricLimitersDev(),
+ para->getParD(level)->isEvenTimestep,
+ indices,
+ size_indices);
+ break;
+
+ case CollisionTemplate::WriteMacroVars:
+ LB_Kernel_CumulantK17 < turbulenceModel, true, false > <<< cudaGrid.grid, cudaGrid.threads, 0, stream >>>( para->getParD(level)->omega,
+ para->getParD(level)->neighborX, para->getParD(level)->neighborY, para->getParD(level)->neighborZ,
+ para->getParD(level)->distributions.f[0],
+ para->getParD(level)->rho,
+ para->getParD(level)->velocityX, para->getParD(level)->velocityY, para->getParD(level)->velocityZ,
+ para->getParD(level)->turbViscosity,
+ para->getSGSConstant(),
+ (unsigned long)para->getParD(level)->numberOfNodes,
+ level,
+ para->getForcesDev(),
+ para->getParD(level)->forceX_SP, para->getParD(level)->forceY_SP, para->getParD(level)->forceZ_SP,
+ para->getQuadricLimitersDev(),
+ para->getParD(level)->isEvenTimestep,
+ indices,
+ size_indices);
+ break;
+
+ case CollisionTemplate::SubDomainBorder:
+ case CollisionTemplate::AllFeatures:
+ LB_Kernel_CumulantK17 < turbulenceModel, true, true > <<< cudaGrid.grid, cudaGrid.threads, 0, stream >>>( para->getParD(level)->omega,
+ para->getParD(level)->neighborX, para->getParD(level)->neighborY, para->getParD(level)->neighborZ,
+ para->getParD(level)->distributions.f[0],
+ para->getParD(level)->rho,
+ para->getParD(level)->velocityX, para->getParD(level)->velocityY, para->getParD(level)->velocityZ,
+ para->getParD(level)->turbViscosity,
+ para->getSGSConstant(),
+ (unsigned long)para->getParD(level)->numberOfNodes,
+ level,
+ para->getForcesDev(),
+ para->getParD(level)->forceX_SP, para->getParD(level)->forceY_SP, para->getParD(level)->forceZ_SP,
+ para->getQuadricLimitersDev(),
+ para->getParD(level)->isEvenTimestep,
+ indices,
+ size_indices);
+ break; case CollisionTemplate::ApplyBodyForce:
+ LB_Kernel_CumulantK17 < turbulenceModel, false, true > <<< cudaGrid.grid, cudaGrid.threads, 0, stream >>>( para->getParD(level)->omega,
+ para->getParD(level)->neighborX, para->getParD(level)->neighborY, para->getParD(level)->neighborZ,
+ para->getParD(level)->distributions.f[0],
+ para->getParD(level)->rho,
+ para->getParD(level)->velocityX, para->getParD(level)->velocityY, para->getParD(level)->velocityZ,
+ para->getParD(level)->turbViscosity,
+ para->getSGSConstant(),
+ (unsigned long)para->getParD(level)->numberOfNodes,
+ level,
+ para->getForcesDev(),
+ para->getParD(level)->forceX_SP, para->getParD(level)->forceY_SP, para->getParD(level)->forceZ_SP,
+ para->getQuadricLimitersDev(),
+ para->getParD(level)->isEvenTimestep,
+ indices,
+ size_indices);
+ break; default:
+ throw std::runtime_error("Invalid CollisionTemplate in CumulantK17::runOnIndices()");
+ break;
+ }
+
+ getLastCudaError("LB_Kernel_CumulantK17 execution failed");
+}
+
+template<TurbulenceModel turbulenceModel>
+CumulantK17<turbulenceModel>::CumulantK17(std::shared_ptr<Parameter> para, int level)
+{
+ this->para = para;
+ this->level = level;
+
+ myPreProcessorTypes.push_back(InitCompSP27);
+
+ myKernelGroup = BasicKernel;
+
+ this->cudaGrid = vf::cuda::CudaGrid(para->getParD(level)->numberofthreads, para->getParD(level)->numberOfNodes);
+ this->kernelUsesFluidNodeIndices = true;
+
+ VF_LOG_INFO("Using turbulence model: {}", turbulenceModel);
+}
+
+template class CumulantK17<TurbulenceModel::AMD>;
+template class CumulantK17<TurbulenceModel::Smagorinsky>;
+template class CumulantK17<TurbulenceModel::QR>;
+template class CumulantK17<TurbulenceModel::None>;
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17.h b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17.h
new file mode 100644
index 0000000000000000000000000000000000000000..00c79a30c9ccf9a89901165d020fc85d5a479c1d
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17.h
@@ -0,0 +1,20 @@
+#ifndef CUMULANT_K17_H
+#define CUMULANT_K17_H
+
+#include "Kernel/KernelImp.h"
+#include "Parameter/Parameter.h"
+
+template<TurbulenceModel turbulenceModel>
+class CumulantK17 : public KernelImp
+{
+public:
+ static std::shared_ptr< CumulantK17<turbulenceModel> > getNewInstance(std::shared_ptr< Parameter> para, int level);
+ void run() override;
+ void runOnIndices(const unsigned int *indices, unsigned int size_indices, CollisionTemplate collisionTemplate, CudaStreamIndex streamIndex) override;
+
+private:
+ CumulantK17();
+ CumulantK17(std::shared_ptr<Parameter> para, int level);
+};
+
+#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp.cu
deleted file mode 100644
index b176b94d07e7f280d738a797d5bd853095e3caed..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp.cu
+++ /dev/null
@@ -1,33 +0,0 @@
-#include "CumulantK17Comp.h"
-
-#include "Parameter/Parameter.h"
-#include "CumulantK17Comp_Device.cuh"
-#include "cuda/CudaGrid.h"
-
-std::shared_ptr<CumulantK17Comp> CumulantK17Comp::getNewInstance(std::shared_ptr<Parameter> para, int level)
-{
- return std::shared_ptr<CumulantK17Comp>(new CumulantK17Comp(para,level));
-}
-
-void CumulantK17Comp::run()
-{
- LB_Kernel_CumulantK17Comp <<< cudaGrid.grid, cudaGrid.threads >>>(para->getParD(level)->omega,
- para->getParD(level)->typeOfGridNode,
- para->getParD(level)->neighborX,
- para->getParD(level)->neighborY,
- para->getParD(level)->neighborZ,
- para->getParD(level)->distributions.f[0],
- para->getParD(level)->numberOfNodes,
- level,
- para->getForcesDev(),
- para->getQuadricLimitersDev(),
- para->getParD(level)->isEvenTimestep);
- getLastCudaError("LB_Kernel_CumulantK17Comp execution failed");
-}
-
-CumulantK17Comp::CumulantK17Comp(std::shared_ptr<Parameter> para, int level): KernelImp(para, level)
-{
- myPreProcessorTypes.push_back(InitCompSP27);
- myKernelGroup = BasicKernel;
- this->cudaGrid = vf::cuda::CudaGrid(para->getParD(level)->numberofthreads, para->getParD(level)->numberOfNodes);
-}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp.h b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp.h
deleted file mode 100644
index 22a95a688e5d078d7b710f494bfea360c9af0d6b..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp.h
+++ /dev/null
@@ -1,17 +0,0 @@
-#ifndef CUMULANT_K17_COMP_H
-#define CUMULANT_K17_COMP_H
-
-#include "Kernel/KernelImp.h"
-
-class CumulantK17Comp : public KernelImp
-{
-public:
- static std::shared_ptr<CumulantK17Comp> getNewInstance(std::shared_ptr< Parameter> para, int level);
- void run();
-
-private:
- CumulantK17Comp();
- CumulantK17Comp(std::shared_ptr< Parameter> para, int level);
-};
-
-#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp_Device.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp_Device.cu
deleted file mode 100644
index 7cf27aa883cbfd3a0e4a0a36fa61649a62d06eeb..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp_Device.cu
+++ /dev/null
@@ -1,1040 +0,0 @@
-#include "LBM/LB.h"
-#include "lbm/constants/D3Q27.h"
-#include <lbm/constants/NumericConstants.h>
-
-using namespace vf::lbm::constant;
-using namespace vf::lbm::dir;
-#include "math.h"
-
-
-__global__ void LB_Kernel_CumulantK17Comp(real omega,
- unsigned int* bcMatD,
- unsigned int* neighborX,
- unsigned int* neighborY,
- unsigned int* neighborZ,
- real* DDStart,
- int size_Mat,
- int level,
- real* forces,
- real* quadricLimiters,
- bool EvenOrOdd)
-{
- ////////////////////////////////////////////////////////////////////////////////
- const unsigned x = threadIdx.x; // Globaler x-Index
- const unsigned y = blockIdx.x; // Globaler y-Index
- const unsigned z = blockIdx.y; // Globaler z-Index
-
- const unsigned nx = blockDim.x;
- const unsigned ny = gridDim.x;
-
- const unsigned k = nx*(ny*z + y) + x;
- //////////////////////////////////////////////////////////////////////////
-
- if (k<size_Mat)
- {
- ////////////////////////////////////////////////////////////////////////////////
- unsigned int BC;
- BC = bcMatD[k];
-
- if ((BC != GEO_SOLID) && (BC != GEO_VOID))
- {
- Distributions27 D;
- if (EvenOrOdd == true)
- {
- D.f[DIR_P00] = &DDStart[DIR_P00 *size_Mat];
- D.f[DIR_M00] = &DDStart[DIR_M00 *size_Mat];
- D.f[DIR_0P0] = &DDStart[DIR_0P0 *size_Mat];
- D.f[DIR_0M0] = &DDStart[DIR_0M0 *size_Mat];
- D.f[DIR_00P] = &DDStart[DIR_00P *size_Mat];
- D.f[DIR_00M] = &DDStart[DIR_00M *size_Mat];
- D.f[DIR_PP0] = &DDStart[DIR_PP0 *size_Mat];
- D.f[DIR_MM0] = &DDStart[DIR_MM0 *size_Mat];
- D.f[DIR_PM0] = &DDStart[DIR_PM0 *size_Mat];
- D.f[DIR_MP0] = &DDStart[DIR_MP0 *size_Mat];
- D.f[DIR_P0P] = &DDStart[DIR_P0P *size_Mat];
- D.f[DIR_M0M] = &DDStart[DIR_M0M *size_Mat];
- D.f[DIR_P0M] = &DDStart[DIR_P0M *size_Mat];
- D.f[DIR_M0P] = &DDStart[DIR_M0P *size_Mat];
- D.f[DIR_0PP] = &DDStart[DIR_0PP *size_Mat];
- D.f[DIR_0MM] = &DDStart[DIR_0MM *size_Mat];
- D.f[DIR_0PM] = &DDStart[DIR_0PM *size_Mat];
- D.f[DIR_0MP] = &DDStart[DIR_0MP *size_Mat];
- D.f[DIR_000] = &DDStart[DIR_000*size_Mat];
- D.f[DIR_PPP] = &DDStart[DIR_PPP *size_Mat];
- D.f[DIR_MMP] = &DDStart[DIR_MMP *size_Mat];
- D.f[DIR_PMP] = &DDStart[DIR_PMP *size_Mat];
- D.f[DIR_MPP] = &DDStart[DIR_MPP *size_Mat];
- D.f[DIR_PPM] = &DDStart[DIR_PPM *size_Mat];
- D.f[DIR_MMM] = &DDStart[DIR_MMM *size_Mat];
- D.f[DIR_PMM] = &DDStart[DIR_PMM *size_Mat];
- D.f[DIR_MPM] = &DDStart[DIR_MPM *size_Mat];
- }
- else
- {
- D.f[DIR_M00] = &DDStart[DIR_P00 *size_Mat];
- D.f[DIR_P00] = &DDStart[DIR_M00 *size_Mat];
- D.f[DIR_0M0] = &DDStart[DIR_0P0 *size_Mat];
- D.f[DIR_0P0] = &DDStart[DIR_0M0 *size_Mat];
- D.f[DIR_00M] = &DDStart[DIR_00P *size_Mat];
- D.f[DIR_00P] = &DDStart[DIR_00M *size_Mat];
- D.f[DIR_MM0] = &DDStart[DIR_PP0 *size_Mat];
- D.f[DIR_PP0] = &DDStart[DIR_MM0 *size_Mat];
- D.f[DIR_MP0] = &DDStart[DIR_PM0 *size_Mat];
- D.f[DIR_PM0] = &DDStart[DIR_MP0 *size_Mat];
- D.f[DIR_M0M] = &DDStart[DIR_P0P *size_Mat];
- D.f[DIR_P0P] = &DDStart[DIR_M0M *size_Mat];
- D.f[DIR_M0P] = &DDStart[DIR_P0M *size_Mat];
- D.f[DIR_P0M] = &DDStart[DIR_M0P *size_Mat];
- D.f[DIR_0MM] = &DDStart[DIR_0PP *size_Mat];
- D.f[DIR_0PP] = &DDStart[DIR_0MM *size_Mat];
- D.f[DIR_0MP] = &DDStart[DIR_0PM *size_Mat];
- D.f[DIR_0PM] = &DDStart[DIR_0MP *size_Mat];
- D.f[DIR_000] = &DDStart[DIR_000*size_Mat];
- D.f[DIR_MMM] = &DDStart[DIR_PPP *size_Mat];
- D.f[DIR_PPM] = &DDStart[DIR_MMP *size_Mat];
- D.f[DIR_MPM] = &DDStart[DIR_PMP *size_Mat];
- D.f[DIR_PMM] = &DDStart[DIR_MPP *size_Mat];
- D.f[DIR_MMP] = &DDStart[DIR_PPM *size_Mat];
- D.f[DIR_PPP] = &DDStart[DIR_MMM *size_Mat];
- D.f[DIR_MPP] = &DDStart[DIR_PMM *size_Mat];
- D.f[DIR_PMP] = &DDStart[DIR_MPM *size_Mat];
- }
-
- ////////////////////////////////////////////////////////////////////////////////
- //index
- //unsigned int kzero= k;
- //unsigned int ke = k;
- unsigned int kw = neighborX[k];
- //unsigned int kn = k;
- unsigned int ks = neighborY[k];
- //unsigned int kt = k;
- unsigned int kb = neighborZ[k];
- unsigned int ksw = neighborY[kw];
- //unsigned int kne = k;
- //unsigned int kse = ks;
- //unsigned int knw = kw;
- unsigned int kbw = neighborZ[kw];
- //unsigned int kte = k;
- //unsigned int kbe = kb;
- //unsigned int ktw = kw;
- unsigned int kbs = neighborZ[ks];
- //unsigned int ktn = k;
- //unsigned int kbn = kb;
- //unsigned int kts = ks;
- //unsigned int ktse = ks;
- //unsigned int kbnw = kbw;
- //unsigned int ktnw = kw;
- //unsigned int kbse = kbs;
- //unsigned int ktsw = ksw;
- //unsigned int kbne = kb;
- //unsigned int ktne = k;
- unsigned int kbsw = neighborZ[ksw];
-
- //unsigned int kzero= k;
- //unsigned int ke = k;
- //unsigned int kw = neighborX[k];
- //unsigned int kn = k;
- //unsigned int ks = neighborY[k];
- //unsigned int kt = k;
- //unsigned int kb = neighborZ[k];
- //unsigned int ksw = neighborY[kw];
- //unsigned int kne = k;
- //unsigned int kse = ks;
- //unsigned int knw = kw;
- //unsigned int kbw = neighborZ[kw];
- //unsigned int kte = k;
- //unsigned int kbe = kb;
- //unsigned int ktw = kw;
- //unsigned int kbs = neighborZ[ks];
- //unsigned int ktn = k;
- //unsigned int kbn = kb;
- //unsigned int kts = ks;
- //unsigned int ktse = ks;
- //unsigned int kbnw = kbw;
- //unsigned int ktnw = kw;
- //unsigned int kbse = kbs;
- //unsigned int ktsw = ksw;
- //unsigned int kbne = kb;
- //unsigned int ktne = k;
- //unsigned int kbsw = neighborZ[ksw];
- //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- real mfcbb = (D.f[DIR_P00])[k];//[ke ];// + c2over27 ;(D.f[DIR_P00 ])[k ];//ke
- real mfabb = (D.f[DIR_M00])[kw];//[kw ];// + c2over27 ;(D.f[DIR_M00 ])[kw ];
- real mfbcb = (D.f[DIR_0P0])[k];//[kn ];// + c2over27 ;(D.f[DIR_0P0 ])[k ];//kn
- real mfbab = (D.f[DIR_0M0])[ks];//[ks ];// + c2over27 ;(D.f[DIR_0M0 ])[ks ];
- real mfbbc = (D.f[DIR_00P])[k];//[kt ];// + c2over27 ;(D.f[DIR_00P ])[k ];//kt
- real mfbba = (D.f[DIR_00M])[kb];//[kb ];// + c2over27 ;(D.f[DIR_00M ])[kb ];
- real mfccb = (D.f[DIR_PP0])[k];//[kne ];// + c1over54 ;(D.f[DIR_PP0 ])[k ];//kne
- real mfaab = (D.f[DIR_MM0])[ksw];//[ksw ];// + c1over54 ;(D.f[DIR_MM0 ])[ksw];
- real mfcab = (D.f[DIR_PM0])[ks];//[kse ];// + c1over54 ;(D.f[DIR_PM0 ])[ks ];//kse
- real mfacb = (D.f[DIR_MP0])[kw];//[knw ];// + c1over54 ;(D.f[DIR_MP0 ])[kw ];//knw
- real mfcbc = (D.f[DIR_P0P])[k];//[kte ];// + c1over54 ;(D.f[DIR_P0P ])[k ];//kte
- real mfaba = (D.f[DIR_M0M])[kbw];//[kbw ];// + c1over54 ;(D.f[DIR_M0M ])[kbw];
- real mfcba = (D.f[DIR_P0M])[kb];//[kbe ];// + c1over54 ;(D.f[DIR_P0M ])[kb ];//kbe
- real mfabc = (D.f[DIR_M0P])[kw];//[ktw ];// + c1over54 ;(D.f[DIR_M0P ])[kw ];//ktw
- real mfbcc = (D.f[DIR_0PP])[k];//[ktn ];// + c1over54 ;(D.f[DIR_0PP ])[k ];//ktn
- real mfbaa = (D.f[DIR_0MM])[kbs];//[kbs ];// + c1over54 ;(D.f[DIR_0MM ])[kbs];
- real mfbca = (D.f[DIR_0PM])[kb];//[kbn ];// + c1over54 ;(D.f[DIR_0PM ])[kb ];//kbn
- real mfbac = (D.f[DIR_0MP])[ks];//[kts ];// + c1over54 ;(D.f[DIR_0MP ])[ks ];//kts
- real mfbbb = (D.f[DIR_000])[k];//[kzero];// + c8over27 ;(D.f[DIR_000])[k ];//kzero
- real mfccc = (D.f[DIR_PPP])[k];//[ktne ];// + c1over216;(D.f[DIR_PPP ])[k ];//ktne
- real mfaac = (D.f[DIR_MMP])[ksw];//[ktsw ];// + c1over216;(D.f[DIR_MMP ])[ksw];//ktsw
- real mfcac = (D.f[DIR_PMP])[ks];//[ktse ];// + c1over216;(D.f[DIR_PMP ])[ks ];//ktse
- real mfacc = (D.f[DIR_MPP])[kw];//[ktnw ];// + c1over216;(D.f[DIR_MPP ])[kw ];//ktnw
- real mfcca = (D.f[DIR_PPM])[kb];//[kbne ];// + c1over216;(D.f[DIR_PPM ])[kb ];//kbne
- real mfaaa = (D.f[DIR_MMM])[kbsw];//[kbsw ];// + c1over216;(D.f[DIR_MMM ])[kbsw];
- real mfcaa = (D.f[DIR_PMM])[kbs];//[kbse ];// + c1over216;(D.f[DIR_PMM ])[kbs];//kbse
- real mfaca = (D.f[DIR_MPM])[kbw];//[kbnw ];// + c1over216;(D.f[DIR_MPM ])[kbw];//kbnw
- ////////////////////////////////////////////////////////////////////////////////////
- real drho = ((((mfccc + mfaaa) + (mfaca + mfcac)) + ((mfacc + mfcaa) + (mfaac + mfcca))) +
- (((mfbac + mfbca) + (mfbaa + mfbcc)) + ((mfabc + mfcba) + (mfaba + mfcbc)) + ((mfacb + mfcab) + (mfaab + mfccb))) +
- ((mfabb + mfcbb) + (mfbab + mfbcb) + (mfbba + mfbbc))) + mfbbb;
-
- real rho = c1o1 + drho;
- ////////////////////////////////////////////////////////////////////////////////////
- //slow
- //real oMdrho = one - ((((mfccc+mfaaa) + (mfaca+mfcac)) + ((mfacc+mfcaa) + (mfaac+mfcca))) +
- // (((mfbac+mfbca) + (mfbaa+mfbcc)) + ((mfabc+mfcba) + (mfaba+mfcbc)) + ((mfacb+mfcab) + (mfaab+mfccb))) +
- // ((mfabb+mfcbb) + (mfbab+mfbcb) + (mfbba+mfbbc)));//fehlt mfbbb
- real vvx = ((((mfccc - mfaaa) + (mfcac - mfaca)) + ((mfcaa - mfacc) + (mfcca - mfaac))) +
- (((mfcba - mfabc) + (mfcbc - mfaba)) + ((mfcab - mfacb) + (mfccb - mfaab))) +
- (mfcbb - mfabb)) / rho;
- real vvy = ((((mfccc - mfaaa) + (mfaca - mfcac)) + ((mfacc - mfcaa) + (mfcca - mfaac))) +
- (((mfbca - mfbac) + (mfbcc - mfbaa)) + ((mfacb - mfcab) + (mfccb - mfaab))) +
- (mfbcb - mfbab)) / rho;
- real vvz = ((((mfccc - mfaaa) + (mfcac - mfaca)) + ((mfacc - mfcaa) + (mfaac - mfcca))) +
- (((mfbac - mfbca) + (mfbcc - mfbaa)) + ((mfabc - mfcba) + (mfcbc - mfaba))) +
- (mfbbc - mfbba)) / rho;
- ////////////////////////////////////////////////////////////////////////////////////
- //the force be with you
- real fx = forces[0] / (pow((double)c2o1, (double)level)); //zero;//0.0032653/(pow(two,level)); //0.000000005;//(two/1600000.0) / 120.0; //
- real fy = forces[1] / (pow((double)c2o1, (double)level)); //zero;
- real fz = forces[2] / (pow((double)c2o1, (double)level)); //zero;
- vvx += fx*c1o2;
- vvy += fy*c1o2;
- vvz += fz*c1o2;
- ////////////////////////////////////////////////////////////////////////////////////
- //real omega = omega_in;
- ////////////////////////////////////////////////////////////////////////////////////
- //fast
- real oMdrho = c1o1; // comp special
- //real oMdrho = one - (mfccc+mfaaa + mfaca+mfcac + mfacc+mfcaa + mfaac+mfcca +
- // mfbac+mfbca + mfbaa+mfbcc + mfabc+mfcba + mfaba+mfcbc + mfacb+mfcab + mfaab+mfccb +
- // mfabb+mfcbb + mfbab+mfbcb + mfbba+mfbbc + mfbbb);//fehlt mfbbb nicht mehr
- //real vvx =mfccc-mfaaa + mfcac-mfaca + mfcaa-mfacc + mfcca-mfaac +
- // mfcba-mfabc + mfcbc-mfaba + mfcab-mfacb + mfccb-mfaab +
- // mfcbb-mfabb;
- //real vvy =mfccc-mfaaa + mfaca-mfcac + mfacc-mfcaa + mfcca-mfaac +
- // mfbca-mfbac + mfbcc-mfbaa + mfacb-mfcab + mfccb-mfaab +
- // mfbcb-mfbab;
- //real vvz =mfccc-mfaaa + mfcac-mfaca + mfacc-mfcaa + mfaac-mfcca +
- // mfbac-mfbca + mfbcc-mfbaa + mfabc-mfcba + mfcbc-mfaba +
- // mfbbc-mfbba;
- ////////////////////////////////////////////////////////////////////////////////////
- // oMdrho assembler style -------> faaaaaastaaaa
- // or much sloooowaaaa ... it dep�ndssssss on sadaku
- real m0, m1, m2;
- //real oMdrho;
- //{
- // oMdrho=mfccc+mfaaa;
- // m0=mfaca+mfcac;
- // m1=mfacc+mfcaa;
- // m2=mfaac+mfcca;
- // oMdrho+=m0;
- // m1+=m2;
- // oMdrho+=m1;
- // m0=mfbac+mfbca;
- // m1=mfbaa+mfbcc;
- // m0+=m1;
- // m1=mfabc+mfcba;
- // m2=mfaba+mfcbc;
- // m1+=m2;
- // m0+=m1;
- // m1=mfacb+mfcab;
- // m2=mfaab+mfccb;
- // m1+=m2;
- // m0+=m1;
- // oMdrho+=m0;
- // m0=mfabb+mfcbb;
- // m1=mfbab+mfbcb;
- // m2=mfbba+mfbbc;
- // m0+=m1+m2;
- // m0+=mfbbb; //hat gefehlt
- // oMdrho = one - (oMdrho + m0);
- //}
- //real vvx;
- real vx2;
- //{
- // vvx = mfccc-mfaaa;
- // m0 = mfcac-mfaca;
- // m1 = mfcaa-mfacc;
- // m2 = mfcca-mfaac;
- // vvx+= m0;
- // m1 += m2;
- // vvx+= m1;
- // vx2 = mfcba-mfabc;
- // m0 = mfcbc-mfaba;
- // m1 = mfcab-mfacb;
- // m2 = mfccb-mfaab;
- // vx2+= m0;
- // m1 += m2;
- // vx2+= m1;
- // vvx+= vx2;
- // vx2 = mfcbb-mfabb;
- // vvx+= vx2;
- //}
- //real vvy;
- real vy2;
- //{
- // vvy = mfccc-mfaaa;
- // m0 = mfaca-mfcac;
- // m1 = mfacc-mfcaa;
- // m2 = mfcca-mfaac;
- // vvy+= m0;
- // m1 += m2;
- // vvy+= m1;
- // vy2 = mfbca-mfbac;
- // m0 = mfbcc-mfbaa;
- // m1 = mfacb-mfcab;
- // m2 = mfccb-mfaab;
- // vy2+= m0;
- // m1 += m2;
- // vy2+= m1;
- // vvy+= vy2;
- // vy2 = mfbcb-mfbab;
- // vvy+= vy2;
- //}
- //real vvz;
- real vz2;
- //{
- // vvz = mfccc-mfaaa;
- // m0 = mfcac-mfaca;
- // m1 = mfacc-mfcaa;
- // m2 = mfaac-mfcca;
- // vvz+= m0;
- // m1 += m2;
- // vvz+= m1;
- // vz2 = mfbac-mfbca;
- // m0 = mfbcc-mfbaa;
- // m1 = mfabc-mfcba;
- // m2 = mfcbc-mfaba;
- // vz2+= m0;
- // m1 += m2;
- // vz2+= m1;
- // vvz+= vz2;
- // vz2 = mfbbc-mfbba;
- // vvz+= vz2;
- //}
- vx2 = vvx*vvx;
- vy2 = vvy*vvy;
- vz2 = vvz*vvz;
- ////////////////////////////////////////////////////////////////////////////////////
- real wadjust;
- real qudricLimitP = quadricLimiters[0]; //0.01f; // * 0.0001f; // 1000000.0f; // 1000000.0f; //
- real qudricLimitM = quadricLimiters[1]; //0.01f; // * 0.0001f; // 1000000.0f; // 1000000.0f; //
- real qudricLimitD = quadricLimiters[2]; //0.01f; // * 0.001f; // 1000000.0f; // 1000000.0f; //
- ////////////////////////////////////////////////////////////////////////////////////
- //Hin
- ////////////////////////////////////////////////////////////////////////////////////
- // mit 1/36, 1/9, 1/36, 1/9, 4/9, 1/9, 1/36, 1/9, 1/36 Konditionieren
- ////////////////////////////////////////////////////////////////////////////////////
- // Z - Dir
- m2 = mfaaa + mfaac;
- m1 = mfaac - mfaaa;
- m0 = m2 + mfaab;
- mfaaa = m0;
- m0 += c1o36 * oMdrho;
- mfaab = m1 - m0 * vvz;
- mfaac = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfaba + mfabc;
- m1 = mfabc - mfaba;
- m0 = m2 + mfabb;
- mfaba = m0;
- m0 += c1o9 * oMdrho;
- mfabb = m1 - m0 * vvz;
- mfabc = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfaca + mfacc;
- m1 = mfacc - mfaca;
- m0 = m2 + mfacb;
- mfaca = m0;
- m0 += c1o36 * oMdrho;
- mfacb = m1 - m0 * vvz;
- mfacc = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfbaa + mfbac;
- m1 = mfbac - mfbaa;
- m0 = m2 + mfbab;
- mfbaa = m0;
- m0 += c1o9 * oMdrho;
- mfbab = m1 - m0 * vvz;
- mfbac = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfbba + mfbbc;
- m1 = mfbbc - mfbba;
- m0 = m2 + mfbbb;
- mfbba = m0;
- m0 += c4o9 * oMdrho;
- mfbbb = m1 - m0 * vvz;
- mfbbc = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfbca + mfbcc;
- m1 = mfbcc - mfbca;
- m0 = m2 + mfbcb;
- mfbca = m0;
- m0 += c1o9 * oMdrho;
- mfbcb = m1 - m0 * vvz;
- mfbcc = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfcaa + mfcac;
- m1 = mfcac - mfcaa;
- m0 = m2 + mfcab;
- mfcaa = m0;
- m0 += c1o36 * oMdrho;
- mfcab = m1 - m0 * vvz;
- mfcac = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfcba + mfcbc;
- m1 = mfcbc - mfcba;
- m0 = m2 + mfcbb;
- mfcba = m0;
- m0 += c1o9 * oMdrho;
- mfcbb = m1 - m0 * vvz;
- mfcbc = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfcca + mfccc;
- m1 = mfccc - mfcca;
- m0 = m2 + mfccb;
- mfcca = m0;
- m0 += c1o36 * oMdrho;
- mfccb = m1 - m0 * vvz;
- mfccc = m2 - c2o1* m1 * vvz + vz2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- // mit 1/6, 0, 1/18, 2/3, 0, 2/9, 1/6, 0, 1/18 Konditionieren
- ////////////////////////////////////////////////////////////////////////////////////
- // Y - Dir
- m2 = mfaaa + mfaca;
- m1 = mfaca - mfaaa;
- m0 = m2 + mfaba;
- mfaaa = m0;
- m0 += c1o6 * oMdrho;
- mfaba = m1 - m0 * vvy;
- mfaca = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfaab + mfacb;
- m1 = mfacb - mfaab;
- m0 = m2 + mfabb;
- mfaab = m0;
- mfabb = m1 - m0 * vvy;
- mfacb = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfaac + mfacc;
- m1 = mfacc - mfaac;
- m0 = m2 + mfabc;
- mfaac = m0;
- m0 += c1o18 * oMdrho;
- mfabc = m1 - m0 * vvy;
- mfacc = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfbaa + mfbca;
- m1 = mfbca - mfbaa;
- m0 = m2 + mfbba;
- mfbaa = m0;
- m0 += c2o3 * oMdrho;
- mfbba = m1 - m0 * vvy;
- mfbca = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfbab + mfbcb;
- m1 = mfbcb - mfbab;
- m0 = m2 + mfbbb;
- mfbab = m0;
- mfbbb = m1 - m0 * vvy;
- mfbcb = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfbac + mfbcc;
- m1 = mfbcc - mfbac;
- m0 = m2 + mfbbc;
- mfbac = m0;
- m0 += c2o9 * oMdrho;
- mfbbc = m1 - m0 * vvy;
- mfbcc = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfcaa + mfcca;
- m1 = mfcca - mfcaa;
- m0 = m2 + mfcba;
- mfcaa = m0;
- m0 += c1o6 * oMdrho;
- mfcba = m1 - m0 * vvy;
- mfcca = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfcab + mfccb;
- m1 = mfccb - mfcab;
- m0 = m2 + mfcbb;
- mfcab = m0;
- mfcbb = m1 - m0 * vvy;
- mfccb = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfcac + mfccc;
- m1 = mfccc - mfcac;
- m0 = m2 + mfcbc;
- mfcac = m0;
- m0 += c1o18 * oMdrho;
- mfcbc = m1 - m0 * vvy;
- mfccc = m2 - c2o1* m1 * vvy + vy2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- // mit 1, 0, 1/3, 0, 0, 0, 1/3, 0, 1/9 Konditionieren
- ////////////////////////////////////////////////////////////////////////////////////
- // X - Dir
- m2 = mfaaa + mfcaa;
- m1 = mfcaa - mfaaa;
- m0 = m2 + mfbaa;
- mfaaa = m0;
- m0 += c1o1* oMdrho;
- mfbaa = m1 - m0 * vvx;
- mfcaa = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfaba + mfcba;
- m1 = mfcba - mfaba;
- m0 = m2 + mfbba;
- mfaba = m0;
- mfbba = m1 - m0 * vvx;
- mfcba = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfaca + mfcca;
- m1 = mfcca - mfaca;
- m0 = m2 + mfbca;
- mfaca = m0;
- m0 += c1o3 * oMdrho;
- mfbca = m1 - m0 * vvx;
- mfcca = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfaab + mfcab;
- m1 = mfcab - mfaab;
- m0 = m2 + mfbab;
- mfaab = m0;
- mfbab = m1 - m0 * vvx;
- mfcab = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfabb + mfcbb;
- m1 = mfcbb - mfabb;
- m0 = m2 + mfbbb;
- mfabb = m0;
- mfbbb = m1 - m0 * vvx;
- mfcbb = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfacb + mfccb;
- m1 = mfccb - mfacb;
- m0 = m2 + mfbcb;
- mfacb = m0;
- mfbcb = m1 - m0 * vvx;
- mfccb = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfaac + mfcac;
- m1 = mfcac - mfaac;
- m0 = m2 + mfbac;
- mfaac = m0;
- m0 += c1o3 * oMdrho;
- mfbac = m1 - m0 * vvx;
- mfcac = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfabc + mfcbc;
- m1 = mfcbc - mfabc;
- m0 = m2 + mfbbc;
- mfabc = m0;
- mfbbc = m1 - m0 * vvx;
- mfcbc = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- m2 = mfacc + mfccc;
- m1 = mfccc - mfacc;
- m0 = m2 + mfbcc;
- mfacc = m0;
- m0 += c1o9 * oMdrho;
- mfbcc = m1 - m0 * vvx;
- mfccc = m2 - c2o1* m1 * vvx + vx2 * m0;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
-
- ////////////////////////////////////////////////////////////////////////////////////
- // Cumulants
- ////////////////////////////////////////////////////////////////////////////////////
- real OxxPyyPzz = c1o1; //set the bulk viscosity one is high / two is very low and zero is (too) high ... (also called omega 2)
-
- ////////////////////////////////////////////////////////////
- //3.
- //////////////////////////////
- real OxyyPxzz = c8o1*(-c2o1 + omega)*(c1o1 + c2o1*omega) / (-c8o1 - c14o1*omega + c7o1*omega*omega);//one;
- real OxyyMxzz = c8o1*(-c2o1 + omega)*(-c7o1 + c4o1*omega) / (c56o1 - c50o1*omega + c9o1*omega*omega);//one;
- real Oxyz = c24o1*(-c2o1 + omega)*(-c2o1 - c7o1*omega + c3o1*omega*omega) / (c48o1 + c152o1*omega - c130o1*omega*omega + c29o1*omega*omega*omega);//one;
- ////////////////////////////////////////////////////////////
- //4.
- //////////////////////////////
- real O4 = c1o1;
- //////////////////////////////
- //real O4 = omega;//TRT
- ////////////////////////////////////////////////////////////
- //5.
- //////////////////////////////
- real O5 = c1o1;
- ////////////////////////////////////////////////////////////
- //6.
- //////////////////////////////
- real O6 = c1o1;
- ////////////////////////////////////////////////////////////
-
-
- //central moments to cumulants
- //4.
- real CUMcbb = mfcbb - ((mfcaa + c1o3) * mfabb + c2o1 * mfbba * mfbab) / rho;
- real CUMbcb = mfbcb - ((mfaca + c1o3) * mfbab + c2o1 * mfbba * mfabb) / rho;
- real CUMbbc = mfbbc - ((mfaac + c1o3) * mfbba + c2o1 * mfbab * mfabb) / rho;
-
- real CUMcca = mfcca - (((mfcaa * mfaca + c2o1 * mfbba * mfbba) + c1o3 * (mfcaa + mfaca)) / rho - c1o9*(drho / rho));
- real CUMcac = mfcac - (((mfcaa * mfaac + c2o1 * mfbab * mfbab) + c1o3 * (mfcaa + mfaac)) / rho - c1o9*(drho / rho));
- real CUMacc = mfacc - (((mfaac * mfaca + c2o1 * mfabb * mfabb) + c1o3 * (mfaac + mfaca)) / rho - c1o9*(drho / rho));
-
- //5.
- real CUMbcc = mfbcc - ((mfaac * mfbca + mfaca * mfbac + c4o1 * mfabb * mfbbb + c2o1 * (mfbab * mfacb + mfbba * mfabc)) + c1o3 * (mfbca + mfbac)) / rho;
- real CUMcbc = mfcbc - ((mfaac * mfcba + mfcaa * mfabc + c4o1 * mfbab * mfbbb + c2o1 * (mfabb * mfcab + mfbba * mfbac)) + c1o3 * (mfcba + mfabc)) / rho;
- real CUMccb = mfccb - ((mfcaa * mfacb + mfaca * mfcab + c4o1 * mfbba * mfbbb + c2o1 * (mfbab * mfbca + mfabb * mfcba)) + c1o3 * (mfacb + mfcab)) / rho;
-
- //6.
-
- real CUMccc = mfccc + ((-c4o1 * mfbbb * mfbbb
- - (mfcaa * mfacc + mfaca * mfcac + mfaac * mfcca)
- - c4o1 * (mfabb * mfcbb + mfbab * mfbcb + mfbba * mfbbc)
- - c2o1 * (mfbca * mfbac + mfcba * mfabc + mfcab * mfacb)) / rho
- + (c4o1 * (mfbab * mfbab * mfaca + mfabb * mfabb * mfcaa + mfbba * mfbba * mfaac)
- + c2o1 * (mfcaa * mfaca * mfaac)
- + c16o1 * mfbba * mfbab * mfabb) / (rho * rho)
- - c1o3 * (mfacc + mfcac + mfcca) / rho
- - c1o9 * (mfcaa + mfaca + mfaac) / rho
- + (c2o1 * (mfbab * mfbab + mfabb * mfabb + mfbba * mfbba)
- + (mfaac * mfaca + mfaac * mfcaa + mfaca * mfcaa) + c1o3 *(mfaac + mfaca + mfcaa)) / (rho * rho) * c2o3
- + c1o27*((drho * drho - drho) / (rho*rho)));
-
- //2.
- // linear combinations
- real mxxPyyPzz = mfcaa + mfaca + mfaac;
- real mxxMyy = mfcaa - mfaca;
- real mxxMzz = mfcaa - mfaac;
-
- ////////////////////////////////////////////////////////////////////////////
- real Dxy = -c3o1*omega*mfbba;
- real Dxz = -c3o1*omega*mfbab;
- real Dyz = -c3o1*omega*mfabb;
-
- //3.
- // linear combinations
-
- real mxxyPyzz = mfcba + mfabc;
- real mxxyMyzz = mfcba - mfabc;
-
- real mxxzPyyz = mfcab + mfacb;
- real mxxzMyyz = mfcab - mfacb;
-
- real mxyyPxzz = mfbca + mfbac;
- real mxyyMxzz = mfbca - mfbac;
-
- ///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- //incl. correction (hat noch nicht so gut funktioniert...Optimierungsbedarf??)
-
- real dxux = c1o2 * (-omega) *(mxxMyy + mxxMzz) + c1o2 * OxxPyyPzz * (mfaaa - mxxPyyPzz);
- real dyuy = dxux + omega * c3o2 * mxxMyy;
- real dzuz = dxux + omega * c3o2 * mxxMzz;
-
- //relax
- mxxPyyPzz += OxxPyyPzz*(mfaaa - mxxPyyPzz) - c3o1 * (c1o1 - c1o2 * OxxPyyPzz) * (vx2 * dxux + vy2 * dyuy + vz2 * dzuz);//-magicBulk*OxxPyyPzz;
- mxxMyy += omega * (-mxxMyy) - c3o1 * (c1o1 + c1o2 * (-omega)) * (vx2 * dxux - vy2 * dyuy);
- mxxMzz += omega * (-mxxMzz) - c3o1 * (c1o1 + c1o2 * (-omega)) * (vx2 * dxux - vz2 * dzuz);
-
- ///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
-
- /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- ////no correction
- //mxxPyyPzz += OxxPyyPzz*(mfaaa-mxxPyyPzz);//-magicBulk*OxxPyyPzz;
- //mxxMyy += -(-omega) * (-mxxMyy);
- //mxxMzz += -(-omega) * (-mxxMzz);
- /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
- mfabb += omega * (-mfabb);
- mfbab += omega * (-mfbab);
- mfbba += omega * (-mfbba);
- //////////////////////////////////////////////////////////////////////////
-
- // linear combinations back
- mfcaa = c1o3 * (mxxMyy + mxxMzz + mxxPyyPzz);
- mfaca = c1o3 * (-c2o1* mxxMyy + mxxMzz + mxxPyyPzz);
- mfaac = c1o3 * (mxxMyy - c2o1* mxxMzz + mxxPyyPzz);
-
-
- //relax
- //////////////////////////////////////////////////////////////////////////
- //das ist der limiter
- wadjust = Oxyz + (c1o1 - Oxyz)*abs(mfbbb) / (abs(mfbbb) + qudricLimitD);
- mfbbb += wadjust * (-mfbbb);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz)*abs(mxxyPyzz) / (abs(mxxyPyzz) + qudricLimitP);
- mxxyPyzz += wadjust * (-mxxyPyzz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz)*abs(mxxyMyzz) / (abs(mxxyMyzz) + qudricLimitM);
- mxxyMyzz += wadjust * (-mxxyMyzz);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz)*abs(mxxzPyyz) / (abs(mxxzPyyz) + qudricLimitP);
- mxxzPyyz += wadjust * (-mxxzPyyz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz)*abs(mxxzMyyz) / (abs(mxxzMyyz) + qudricLimitM);
- mxxzMyyz += wadjust * (-mxxzMyyz);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz)*abs(mxyyPxzz) / (abs(mxyyPxzz) + qudricLimitP);
- mxyyPxzz += wadjust * (-mxyyPxzz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz)*abs(mxyyMxzz) / (abs(mxyyMxzz) + qudricLimitM);
- mxyyMxzz += wadjust * (-mxyyMxzz);
- //////////////////////////////////////////////////////////////////////////
- //ohne limiter
- //mfbbb += OxyyMxzz * (-mfbbb);
- //mxxyPyzz += OxyyPxzz * (-mxxyPyzz);
- //mxxyMyzz += OxyyMxzz * (-mxxyMyzz);
- //mxxzPyyz += OxyyPxzz * (-mxxzPyyz);
- //mxxzMyyz += OxyyMxzz * (-mxxzMyyz);
- //mxyyPxzz += OxyyPxzz * (-mxyyPxzz);
- //mxyyMxzz += OxyyMxzz * (-mxyyMxzz);
- //////////////////////////////////////////////////////////////////////////
-
- // linear combinations back
- mfcba = (mxxyMyzz + mxxyPyzz) * c1o2;
- mfabc = (-mxxyMyzz + mxxyPyzz) * c1o2;
- mfcab = (mxxzMyyz + mxxzPyyz) * c1o2;
- mfacb = (-mxxzMyyz + mxxzPyyz) * c1o2;
- mfbca = (mxyyMxzz + mxyyPxzz) * c1o2;
- mfbac = (-mxyyMxzz + mxyyPxzz) * c1o2;
-
- //4.
- //////////////////////////////////////////////////////////////////////////
- //mit limiter
- // wadjust = O4+(one-O4)*abs(CUMacc)/(abs(CUMacc)+qudricLimit);
- //CUMacc += wadjust * (-CUMacc);
- // wadjust = O4+(one-O4)*abs(CUMcac)/(abs(CUMcac)+qudricLimit);
- //CUMcac += wadjust * (-CUMcac);
- // wadjust = O4+(one-O4)*abs(CUMcca)/(abs(CUMcca)+qudricLimit);
- //CUMcca += wadjust * (-CUMcca);
-
- // wadjust = O4+(one-O4)*abs(CUMbbc)/(abs(CUMbbc)+qudricLimit);
- //CUMbbc += wadjust * (-CUMbbc);
- // wadjust = O4+(one-O4)*abs(CUMbcb)/(abs(CUMbcb)+qudricLimit);
- //CUMbcb += wadjust * (-CUMbcb);
- // wadjust = O4+(one-O4)*abs(CUMcbb)/(abs(CUMcbb)+qudricLimit);
- //CUMcbb += wadjust * (-CUMcbb);
- //////////////////////////////////////////////////////////////////////////
- real factorA = (c4o1 + c2o1*omega - c3o1*omega*omega) / (c2o1 - c7o1*omega + c5o1*omega*omega);
- real factorB = (c4o1 + c28o1*omega - c14o1*omega*omega) / (c6o1 - c21o1*omega + c15o1*omega*omega);
- //////////////////////////////////////////////////////////////////////////
- //ohne limiter
- //CUMacc += O4 * (-CUMacc);
- //CUMcac += O4 * (-CUMcac);
- //CUMcca += O4 * (-CUMcca);
- //CUMbbc += O4 * (-CUMbbc);
- //CUMbcb += O4 * (-CUMbcb);
- //CUMcbb += O4 * (-CUMcbb);
- CUMacc = -O4*(c1o1 / omega - c1o2) * (dyuy + dzuz) * c2o3 * factorA + (c1o1 - O4) * (CUMacc);
- CUMcac = -O4*(c1o1 / omega - c1o2) * (dxux + dzuz) * c2o3 * factorA + (c1o1 - O4) * (CUMcac);
- CUMcca = -O4*(c1o1 / omega - c1o2) * (dyuy + dxux) * c2o3 * factorA + (c1o1 - O4) * (CUMcca);
- CUMbbc = -O4*(c1o1 / omega - c1o2) * Dxy * c1o3 * factorB + (c1o1 - O4) * (CUMbbc);
- CUMbcb = -O4*(c1o1 / omega - c1o2) * Dxz * c1o3 * factorB + (c1o1 - O4) * (CUMbcb);
- CUMcbb = -O4*(c1o1 / omega - c1o2) * Dyz * c1o3 * factorB + (c1o1 - O4) * (CUMcbb);
- //////////////////////////////////////////////////////////////////////////
-
-
- //5.
- CUMbcc += O5 * (-CUMbcc);
- CUMcbc += O5 * (-CUMcbc);
- CUMccb += O5 * (-CUMccb);
-
- //6.
- CUMccc += O6 * (-CUMccc);
-
-
-
- //back cumulants to central moments
- //4.
- mfcbb = CUMcbb + ((mfcaa + c1o3) * mfabb + c2o1 * mfbba * mfbab) / rho;
- mfbcb = CUMbcb + ((mfaca + c1o3) * mfbab + c2o1 * mfbba * mfabb) / rho;
- mfbbc = CUMbbc + ((mfaac + c1o3) * mfbba + c2o1 * mfbab * mfabb) / rho;
-
- mfcca = CUMcca + (((mfcaa * mfaca + c2o1 * mfbba * mfbba) + c1o3 * (mfcaa + mfaca)) / rho - c1o9*(drho / rho));
- mfcac = CUMcac + (((mfcaa * mfaac + c2o1 * mfbab * mfbab) + c1o3 * (mfcaa + mfaac)) / rho - c1o9*(drho / rho));
- mfacc = CUMacc + (((mfaac * mfaca + c2o1 * mfabb * mfabb) + c1o3 * (mfaac + mfaca)) / rho - c1o9*(drho / rho));
-
- //5.
- mfbcc = CUMbcc + ((mfaac * mfbca + mfaca * mfbac + c4o1 * mfabb * mfbbb + c2o1 * (mfbab * mfacb + mfbba * mfabc)) + c1o3 * (mfbca + mfbac)) / rho;
- mfcbc = CUMcbc + ((mfaac * mfcba + mfcaa * mfabc + c4o1 * mfbab * mfbbb + c2o1 * (mfabb * mfcab + mfbba * mfbac)) + c1o3 * (mfcba + mfabc)) / rho;
- mfccb = CUMccb + ((mfcaa * mfacb + mfaca * mfcab + c4o1 * mfbba * mfbbb + c2o1 * (mfbab * mfbca + mfabb * mfcba)) + c1o3 * (mfacb + mfcab)) / rho;
-
- //6.
-
- mfccc = CUMccc - ((-c4o1 * mfbbb * mfbbb
- - (mfcaa * mfacc + mfaca * mfcac + mfaac * mfcca)
- - c4o1 * (mfabb * mfcbb + mfbab * mfbcb + mfbba * mfbbc)
- - c2o1 * (mfbca * mfbac + mfcba * mfabc + mfcab * mfacb)) / rho
- + (c4o1 * (mfbab * mfbab * mfaca + mfabb * mfabb * mfcaa + mfbba * mfbba * mfaac)
- + c2o1 * (mfcaa * mfaca * mfaac)
- + c16o1 * mfbba * mfbab * mfabb) / (rho * rho)
- - c1o3 * (mfacc + mfcac + mfcca) / rho
- - c1o9 * (mfcaa + mfaca + mfaac) / rho
- + (c2o1 * (mfbab * mfbab + mfabb * mfabb + mfbba * mfbba)
- + (mfaac * mfaca + mfaac * mfcaa + mfaca * mfcaa) + c1o3 *(mfaac + mfaca + mfcaa)) / (rho * rho) * c2o3
- + c1o27*((drho * drho - drho) / (rho*rho)));
- ////////////////////////////////////////////////////////////////////////////////////
-
- ////////////////////////////////////////////////////////////////////////////////////
- //the force be with you
- mfbaa = -mfbaa;
- mfaba = -mfaba;
- mfaab = -mfaab;
- ////////////////////////////////////////////////////////////////////////////////////
-
-
- ////////////////////////////////////////////////////////////////////////////////////
- //back
- ////////////////////////////////////////////////////////////////////////////////////
- //mit 1, 0, 1/3, 0, 0, 0, 1/3, 0, 1/9 Konditionieren
- ////////////////////////////////////////////////////////////////////////////////////
- // Z - Dir
- m0 = mfaac * c1o2 + mfaab * (vvz - c1o2) + (mfaaa + c1o1* oMdrho) * (vz2 - vvz) * c1o2;
- m1 = -mfaac - c2o1* mfaab * vvz + mfaaa * (c1o1 - vz2) - c1o1* oMdrho * vz2;
- m2 = mfaac * c1o2 + mfaab * (vvz + c1o2) + (mfaaa + c1o1* oMdrho) * (vz2 + vvz) * c1o2;
- mfaaa = m0;
- mfaab = m1;
- mfaac = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfabc * c1o2 + mfabb * (vvz - c1o2) + mfaba * (vz2 - vvz) * c1o2;
- m1 = -mfabc - c2o1* mfabb * vvz + mfaba * (c1o1 - vz2);
- m2 = mfabc * c1o2 + mfabb * (vvz + c1o2) + mfaba * (vz2 + vvz) * c1o2;
- mfaba = m0;
- mfabb = m1;
- mfabc = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfacc * c1o2 + mfacb * (vvz - c1o2) + (mfaca + c1o3 * oMdrho) * (vz2 - vvz) * c1o2;
- m1 = -mfacc - c2o1* mfacb * vvz + mfaca * (c1o1 - vz2) - c1o3 * oMdrho * vz2;
- m2 = mfacc * c1o2 + mfacb * (vvz + c1o2) + (mfaca + c1o3 * oMdrho) * (vz2 + vvz) * c1o2;
- mfaca = m0;
- mfacb = m1;
- mfacc = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfbac * c1o2 + mfbab * (vvz - c1o2) + mfbaa * (vz2 - vvz) * c1o2;
- m1 = -mfbac - c2o1* mfbab * vvz + mfbaa * (c1o1 - vz2);
- m2 = mfbac * c1o2 + mfbab * (vvz + c1o2) + mfbaa * (vz2 + vvz) * c1o2;
- mfbaa = m0;
- mfbab = m1;
- mfbac = m2;
- /////////b//////////////////////////////////////////////////////////////////////////
- m0 = mfbbc * c1o2 + mfbbb * (vvz - c1o2) + mfbba * (vz2 - vvz) * c1o2;
- m1 = -mfbbc - c2o1* mfbbb * vvz + mfbba * (c1o1 - vz2);
- m2 = mfbbc * c1o2 + mfbbb * (vvz + c1o2) + mfbba * (vz2 + vvz) * c1o2;
- mfbba = m0;
- mfbbb = m1;
- mfbbc = m2;
- /////////b//////////////////////////////////////////////////////////////////////////
- m0 = mfbcc * c1o2 + mfbcb * (vvz - c1o2) + mfbca * (vz2 - vvz) * c1o2;
- m1 = -mfbcc - c2o1* mfbcb * vvz + mfbca * (c1o1 - vz2);
- m2 = mfbcc * c1o2 + mfbcb * (vvz + c1o2) + mfbca * (vz2 + vvz) * c1o2;
- mfbca = m0;
- mfbcb = m1;
- mfbcc = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfcac * c1o2 + mfcab * (vvz - c1o2) + (mfcaa + c1o3 * oMdrho) * (vz2 - vvz) * c1o2;
- m1 = -mfcac - c2o1* mfcab * vvz + mfcaa * (c1o1 - vz2) - c1o3 * oMdrho * vz2;
- m2 = mfcac * c1o2 + mfcab * (vvz + c1o2) + (mfcaa + c1o3 * oMdrho) * (vz2 + vvz) * c1o2;
- mfcaa = m0;
- mfcab = m1;
- mfcac = m2;
- /////////c//////////////////////////////////////////////////////////////////////////
- m0 = mfcbc * c1o2 + mfcbb * (vvz - c1o2) + mfcba * (vz2 - vvz) * c1o2;
- m1 = -mfcbc - c2o1* mfcbb * vvz + mfcba * (c1o1 - vz2);
- m2 = mfcbc * c1o2 + mfcbb * (vvz + c1o2) + mfcba * (vz2 + vvz) * c1o2;
- mfcba = m0;
- mfcbb = m1;
- mfcbc = m2;
- /////////c//////////////////////////////////////////////////////////////////////////
- m0 = mfccc * c1o2 + mfccb * (vvz - c1o2) + (mfcca + c1o9 * oMdrho) * (vz2 - vvz) * c1o2;
- m1 = -mfccc - c2o1* mfccb * vvz + mfcca * (c1o1 - vz2) - c1o9 * oMdrho * vz2;
- m2 = mfccc * c1o2 + mfccb * (vvz + c1o2) + (mfcca + c1o9 * oMdrho) * (vz2 + vvz) * c1o2;
- mfcca = m0;
- mfccb = m1;
- mfccc = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- //mit 1/6, 2/3, 1/6, 0, 0, 0, 1/18, 2/9, 1/18 Konditionieren
- ////////////////////////////////////////////////////////////////////////////////////
- // Y - Dir
- m0 = mfaca * c1o2 + mfaba * (vvy - c1o2) + (mfaaa + c1o6 * oMdrho) * (vy2 - vvy) * c1o2;
- m1 = -mfaca - c2o1* mfaba * vvy + mfaaa * (c1o1 - vy2) - c1o6 * oMdrho * vy2;
- m2 = mfaca * c1o2 + mfaba * (vvy + c1o2) + (mfaaa + c1o6 * oMdrho) * (vy2 + vvy) * c1o2;
- mfaaa = m0;
- mfaba = m1;
- mfaca = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfacb * c1o2 + mfabb * (vvy - c1o2) + (mfaab + c2o3 * oMdrho) * (vy2 - vvy) * c1o2;
- m1 = -mfacb - c2o1* mfabb * vvy + mfaab * (c1o1 - vy2) - c2o3 * oMdrho * vy2;
- m2 = mfacb * c1o2 + mfabb * (vvy + c1o2) + (mfaab + c2o3 * oMdrho) * (vy2 + vvy) * c1o2;
- mfaab = m0;
- mfabb = m1;
- mfacb = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfacc * c1o2 + mfabc * (vvy - c1o2) + (mfaac + c1o6 * oMdrho) * (vy2 - vvy) * c1o2;
- m1 = -mfacc - c2o1* mfabc * vvy + mfaac * (c1o1 - vy2) - c1o6 * oMdrho * vy2;
- m2 = mfacc * c1o2 + mfabc * (vvy + c1o2) + (mfaac + c1o6 * oMdrho) * (vy2 + vvy) * c1o2;
- mfaac = m0;
- mfabc = m1;
- mfacc = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfbca * c1o2 + mfbba * (vvy - c1o2) + mfbaa * (vy2 - vvy) * c1o2;
- m1 = -mfbca - c2o1* mfbba * vvy + mfbaa * (c1o1 - vy2);
- m2 = mfbca * c1o2 + mfbba * (vvy + c1o2) + mfbaa * (vy2 + vvy) * c1o2;
- mfbaa = m0;
- mfbba = m1;
- mfbca = m2;
- /////////b//////////////////////////////////////////////////////////////////////////
- m0 = mfbcb * c1o2 + mfbbb * (vvy - c1o2) + mfbab * (vy2 - vvy) * c1o2;
- m1 = -mfbcb - c2o1* mfbbb * vvy + mfbab * (c1o1 - vy2);
- m2 = mfbcb * c1o2 + mfbbb * (vvy + c1o2) + mfbab * (vy2 + vvy) * c1o2;
- mfbab = m0;
- mfbbb = m1;
- mfbcb = m2;
- /////////b//////////////////////////////////////////////////////////////////////////
- m0 = mfbcc * c1o2 + mfbbc * (vvy - c1o2) + mfbac * (vy2 - vvy) * c1o2;
- m1 = -mfbcc - c2o1* mfbbc * vvy + mfbac * (c1o1 - vy2);
- m2 = mfbcc * c1o2 + mfbbc * (vvy + c1o2) + mfbac * (vy2 + vvy) * c1o2;
- mfbac = m0;
- mfbbc = m1;
- mfbcc = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfcca * c1o2 + mfcba * (vvy - c1o2) + (mfcaa + c1o18 * oMdrho) * (vy2 - vvy) * c1o2;
- m1 = -mfcca - c2o1* mfcba * vvy + mfcaa * (c1o1 - vy2) - c1o18 * oMdrho * vy2;
- m2 = mfcca * c1o2 + mfcba * (vvy + c1o2) + (mfcaa + c1o18 * oMdrho) * (vy2 + vvy) * c1o2;
- mfcaa = m0;
- mfcba = m1;
- mfcca = m2;
- /////////c//////////////////////////////////////////////////////////////////////////
- m0 = mfccb * c1o2 + mfcbb * (vvy - c1o2) + (mfcab + c2o9 * oMdrho) * (vy2 - vvy) * c1o2;
- m1 = -mfccb - c2o1* mfcbb * vvy + mfcab * (c1o1 - vy2) - c2o9 * oMdrho * vy2;
- m2 = mfccb * c1o2 + mfcbb * (vvy + c1o2) + (mfcab + c2o9 * oMdrho) * (vy2 + vvy) * c1o2;
- mfcab = m0;
- mfcbb = m1;
- mfccb = m2;
- /////////c//////////////////////////////////////////////////////////////////////////
- m0 = mfccc * c1o2 + mfcbc * (vvy - c1o2) + (mfcac + c1o18 * oMdrho) * (vy2 - vvy) * c1o2;
- m1 = -mfccc - c2o1* mfcbc * vvy + mfcac * (c1o1 - vy2) - c1o18 * oMdrho * vy2;
- m2 = mfccc * c1o2 + mfcbc * (vvy + c1o2) + (mfcac + c1o18 * oMdrho) * (vy2 + vvy) * c1o2;
- mfcac = m0;
- mfcbc = m1;
- mfccc = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- //mit 1/36, 1/9, 1/36, 1/9, 4/9, 1/9, 1/36, 1/9, 1/36 Konditionieren
- ////////////////////////////////////////////////////////////////////////////////////
- // X - Dir
- m0 = mfcaa * c1o2 + mfbaa * (vvx - c1o2) + (mfaaa + c1o36 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfcaa - c2o1* mfbaa * vvx + mfaaa * (c1o1 - vx2) - c1o36 * oMdrho * vx2;
- m2 = mfcaa * c1o2 + mfbaa * (vvx + c1o2) + (mfaaa + c1o36 * oMdrho) * (vx2 + vvx) * c1o2;
- mfaaa = m0;
- mfbaa = m1;
- mfcaa = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfcba * c1o2 + mfbba * (vvx - c1o2) + (mfaba + c1o9 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfcba - c2o1* mfbba * vvx + mfaba * (c1o1 - vx2) - c1o9 * oMdrho * vx2;
- m2 = mfcba * c1o2 + mfbba * (vvx + c1o2) + (mfaba + c1o9 * oMdrho) * (vx2 + vvx) * c1o2;
- mfaba = m0;
- mfbba = m1;
- mfcba = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfcca * c1o2 + mfbca * (vvx - c1o2) + (mfaca + c1o36 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfcca - c2o1* mfbca * vvx + mfaca * (c1o1 - vx2) - c1o36 * oMdrho * vx2;
- m2 = mfcca * c1o2 + mfbca * (vvx + c1o2) + (mfaca + c1o36 * oMdrho) * (vx2 + vvx) * c1o2;
- mfaca = m0;
- mfbca = m1;
- mfcca = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfcab * c1o2 + mfbab * (vvx - c1o2) + (mfaab + c1o9 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfcab - c2o1* mfbab * vvx + mfaab * (c1o1 - vx2) - c1o9 * oMdrho * vx2;
- m2 = mfcab * c1o2 + mfbab * (vvx + c1o2) + (mfaab + c1o9 * oMdrho) * (vx2 + vvx) * c1o2;
- mfaab = m0;
- mfbab = m1;
- mfcab = m2;
- ///////////b////////////////////////////////////////////////////////////////////////
- m0 = mfcbb * c1o2 + mfbbb * (vvx - c1o2) + (mfabb + c4o9 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfcbb - c2o1* mfbbb * vvx + mfabb * (c1o1 - vx2) - c4o9 * oMdrho * vx2;
- m2 = mfcbb * c1o2 + mfbbb * (vvx + c1o2) + (mfabb + c4o9 * oMdrho) * (vx2 + vvx) * c1o2;
- mfabb = m0;
- mfbbb = m1;
- mfcbb = m2;
- ///////////b////////////////////////////////////////////////////////////////////////
- m0 = mfccb * c1o2 + mfbcb * (vvx - c1o2) + (mfacb + c1o9 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfccb - c2o1* mfbcb * vvx + mfacb * (c1o1 - vx2) - c1o9 * oMdrho * vx2;
- m2 = mfccb * c1o2 + mfbcb * (vvx + c1o2) + (mfacb + c1o9 * oMdrho) * (vx2 + vvx) * c1o2;
- mfacb = m0;
- mfbcb = m1;
- mfccb = m2;
- ////////////////////////////////////////////////////////////////////////////////////
- ////////////////////////////////////////////////////////////////////////////////////
- m0 = mfcac * c1o2 + mfbac * (vvx - c1o2) + (mfaac + c1o36 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfcac - c2o1* mfbac * vvx + mfaac * (c1o1 - vx2) - c1o36 * oMdrho * vx2;
- m2 = mfcac * c1o2 + mfbac * (vvx + c1o2) + (mfaac + c1o36 * oMdrho) * (vx2 + vvx) * c1o2;
- mfaac = m0;
- mfbac = m1;
- mfcac = m2;
- ///////////c////////////////////////////////////////////////////////////////////////
- m0 = mfcbc * c1o2 + mfbbc * (vvx - c1o2) + (mfabc + c1o9 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfcbc - c2o1* mfbbc * vvx + mfabc * (c1o1 - vx2) - c1o9 * oMdrho * vx2;
- m2 = mfcbc * c1o2 + mfbbc * (vvx + c1o2) + (mfabc + c1o9 * oMdrho) * (vx2 + vvx) * c1o2;
- mfabc = m0;
- mfbbc = m1;
- mfcbc = m2;
- ///////////c////////////////////////////////////////////////////////////////////////
- m0 = mfccc * c1o2 + mfbcc * (vvx - c1o2) + (mfacc + c1o36 * oMdrho) * (vx2 - vvx) * c1o2;
- m1 = -mfccc - c2o1* mfbcc * vvx + mfacc * (c1o1 - vx2) - c1o36 * oMdrho * vx2;
- m2 = mfccc * c1o2 + mfbcc * (vvx + c1o2) + (mfacc + c1o36 * oMdrho) * (vx2 + vvx) * c1o2;
- mfacc = m0;
- mfbcc = m1;
- mfccc = m2;
- ////////////////////////////////////////////////////////////////////////////////////
-
- ////////////////////////////////////////////////////////////////////////////////////
- (D.f[DIR_P00])[k] = mfabb;//(D.f[ DIR_P00 ])[ke ] = mfabb;// - c2over27 ; (D.f[ DIR_P00 ])[k ]
- (D.f[DIR_M00])[kw] = mfcbb;//(D.f[ DIR_M00 ])[kw ] = mfcbb;// - c2over27 ; (D.f[ DIR_M00 ])[kw ]
- (D.f[DIR_0P0])[k] = mfbab;//(D.f[ DIR_0P0 ])[kn ] = mfbab;// - c2over27 ; (D.f[ DIR_0P0 ])[k ]
- (D.f[DIR_0M0])[ks] = mfbcb;//(D.f[ DIR_0M0 ])[ks ] = mfbcb;// - c2over27 ; (D.f[ DIR_0M0 ])[ks ]
- (D.f[DIR_00P])[k] = mfbba;//(D.f[ DIR_00P ])[kt ] = mfbba;// - c2over27 ; (D.f[ DIR_00P ])[k ]
- (D.f[DIR_00M])[kb] = mfbbc;//(D.f[ DIR_00M ])[kb ] = mfbbc;// - c2over27 ; (D.f[ DIR_00M ])[kb ]
- (D.f[DIR_PP0])[k] = mfaab;//(D.f[ DIR_PP0 ])[kne ] = mfaab;// - c1over54 ; (D.f[ DIR_PP0 ])[k ]
- (D.f[DIR_MM0])[ksw] = mfccb;//(D.f[ DIR_MM0 ])[ksw ] = mfccb;// - c1over54 ; (D.f[ DIR_MM0 ])[ksw ]
- (D.f[DIR_PM0])[ks] = mfacb;//(D.f[ DIR_PM0 ])[kse ] = mfacb;// - c1over54 ; (D.f[ DIR_PM0 ])[ks ]
- (D.f[DIR_MP0])[kw] = mfcab;//(D.f[ DIR_MP0 ])[knw ] = mfcab;// - c1over54 ; (D.f[ DIR_MP0 ])[kw ]
- (D.f[DIR_P0P])[k] = mfaba;//(D.f[ DIR_P0P ])[kte ] = mfaba;// - c1over54 ; (D.f[ DIR_P0P ])[k ]
- (D.f[DIR_M0M])[kbw] = mfcbc;//(D.f[ DIR_M0M ])[kbw ] = mfcbc;// - c1over54 ; (D.f[ DIR_M0M ])[kbw ]
- (D.f[DIR_P0M])[kb] = mfabc;//(D.f[ DIR_P0M ])[kbe ] = mfabc;// - c1over54 ; (D.f[ DIR_P0M ])[kb ]
- (D.f[DIR_M0P])[kw] = mfcba;//(D.f[ DIR_M0P ])[ktw ] = mfcba;// - c1over54 ; (D.f[ DIR_M0P ])[kw ]
- (D.f[DIR_0PP])[k] = mfbaa;//(D.f[ DIR_0PP ])[ktn ] = mfbaa;// - c1over54 ; (D.f[ DIR_0PP ])[k ]
- (D.f[DIR_0MM])[kbs] = mfbcc;//(D.f[ DIR_0MM ])[kbs ] = mfbcc;// - c1over54 ; (D.f[ DIR_0MM ])[kbs ]
- (D.f[DIR_0PM])[kb] = mfbac;//(D.f[ DIR_0PM ])[kbn ] = mfbac;// - c1over54 ; (D.f[ DIR_0PM ])[kb ]
- (D.f[DIR_0MP])[ks] = mfbca;//(D.f[ DIR_0MP ])[kts ] = mfbca;// - c1over54 ; (D.f[ DIR_0MP ])[ks ]
- (D.f[DIR_000])[k] = mfbbb;//(D.f[ DIR_000])[kzero] = mfbbb;// - c8over27 ; (D.f[ DIR_000])[k ]
- (D.f[DIR_PPP])[k] = mfaaa;//(D.f[ DIR_PPP ])[ktne ] = mfaaa;// - c1over216; (D.f[ DIR_PPP ])[k ]
- (D.f[DIR_PMP])[ks] = mfaca;//(D.f[ DIR_PMP ])[ktse ] = mfaca;// - c1over216; (D.f[ DIR_PMP ])[ks ]
- (D.f[DIR_PPM])[kb] = mfaac;//(D.f[ DIR_PPM ])[kbne ] = mfaac;// - c1over216; (D.f[ DIR_PPM ])[kb ]
- (D.f[DIR_PMM])[kbs] = mfacc;//(D.f[ DIR_PMM ])[kbse ] = mfacc;// - c1over216; (D.f[ DIR_PMM ])[kbs ]
- (D.f[DIR_MPP])[kw] = mfcaa;//(D.f[ DIR_MPP ])[ktnw ] = mfcaa;// - c1over216; (D.f[ DIR_MPP ])[kw ]
- (D.f[DIR_MMP])[ksw] = mfcca;//(D.f[ DIR_MMP ])[ktsw ] = mfcca;// - c1over216; (D.f[ DIR_MMP ])[ksw ]
- (D.f[DIR_MPM])[kbw] = mfcac;//(D.f[ DIR_MPM ])[kbnw ] = mfcac;// - c1over216; (D.f[ DIR_MPM ])[kbw ]
- (D.f[DIR_MMM])[kbsw] = mfccc;//(D.f[ DIR_MMM ])[kbsw ] = mfccc;// - c1over216; (D.f[ DIR_MMM ])[kbsw]
- ////////////////////////////////////////////////////////////////////////////////////
- }
- }
-}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp_Device.cuh b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp_Device.cuh
deleted file mode 100644
index f44842057d554498b0b5d4c733e2425e524a3b75..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp_Device.cuh
+++ /dev/null
@@ -1,18 +0,0 @@
-#ifndef LB_Kernel_CUMULANT_K17_COMP_H
-#define LB_Kernel_CUMULANT_K17_COMP_H
-
-#include <DataTypes.h>
-#include <curand.h>
-
-__global__ void LB_Kernel_CumulantK17Comp( real omega,
- unsigned int* bcMatD,
- unsigned int* neighborX,
- unsigned int* neighborY,
- unsigned int* neighborZ,
- real* DDStart,
- int size_Mat,
- int level,
- real* forces,
- real* quadricLimiters,
- bool EvenOrOdd);
-#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesignedDevice.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17_Device.cu
similarity index 65%
rename from src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesignedDevice.cu
rename to src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17_Device.cu
index db8caf1b23c2087a4c5c76886fb4530bc6272a1d..34a444230019a1c6cfacb18e12fd73607fb2e09a 100644
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesignedDevice.cu
+++ b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17_Device.cu
@@ -1,4 +1,4 @@
-//=======================================================================================
+
// ____ ____ __ ______ __________ __ __ __ __
// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
// \ \ | | | | | |_) | | | | | | | / \ | |
@@ -26,9 +26,20 @@
// You should have received a copy of the GNU General Public License along
// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
//
-//! \file Cumulant27chimStream.cu
-//! \ingroup GPU
-//! \author Martin Schoenherr, Anna Wellmann
+//! \file CumlantK17_Device.cu
+//! \author Anna Wellmann, Martin Schönherr, Henry Korb, Henrik Asmuth
+//! \date 05/12/2022
+//! \brief Kernel for CumulantK17 including different turbulence models and options for local body forces and writing macroscopic variables
+//!
+//! CumulantK17 kernel using chimera transformations and quartic limiters as present in Geier et al. (2017). Additional options are three different
+//! eddy-viscosity turbulence models (Smagorinsky, AMD, QR) that can be set via the template parameter turbulenceModel (with default
+//! TurbulenceModel::None).
+//! The kernel is executed separately for each subset of fluid node indices with a different tag CollisionTemplate. For each subset, only the locally
+//! required options are switched on ( \param writeMacroscopicVariables and/or \param applyBodyForce) in order to minimize memory accesses. The default
+//! refers to the plain cumlant kernel (CollisionTemplate::Default).
+//! Nodes are added to subsets (taggedFluidNodes) in Simulation::init using a corresponding tag with different values of CollisionTemplate. These subsets
+//! are provided by the utilized PostCollisionInteractiors depending on they specifc requirements (e.g. writeMacroscopicVariables for probes).
+
//=======================================================================================
/* Device code */
#include "LBM/LB.h"
@@ -36,22 +47,35 @@
#include <lbm/constants/NumericConstants.h>
#include "Kernel/Utilities/DistributionHelper.cuh"
+#include "GPU/TurbulentViscosityInlines.cuh"
+
using namespace vf::lbm::constant;
using namespace vf::lbm::dir;
#include "Kernel/Utilities/ChimeraTransformation.h"
+
////////////////////////////////////////////////////////////////////////////////
-__global__ void LB_Kernel_CumulantK17CompChimRedesigned(
- real omega,
- uint* neighborX,
- uint* neighborY,
- uint* neighborZ,
- real* distributions,
- unsigned long numberOfLBnodes,
- int level,
- real* forces,
- real* quadricLimiters,
- bool isEvenTimestep,
+template<TurbulenceModel turbulenceModel, bool writeMacroscopicVariables, bool applyBodyForce>
+__global__ void LB_Kernel_CumulantK17(
+ real omega_in,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ real* distributions,
+ real* rho,
+ real* vx,
+ real* vy,
+ real* vz,
+ real* turbulentViscosity,
+ real SGSconstant,
+ unsigned long numberOfLBnodes,
+ int level,
+ real* forces,
+ real* bodyForceX,
+ real* bodyForceY,
+ real* bodyForceZ,
+ real* quadricLimiters,
+ bool isEvenTimestep,
const uint *fluidNodeIndices,
uint numberOfFluidNodes)
{
@@ -64,15 +88,14 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
//! The cumulant kernel is executed in the following steps
//!
////////////////////////////////////////////////////////////////////////////////
- //! - Get the thread index coordinates from threadIdx, blockIdx, blockDim and gridDim.
+ //! - Get node index coordinates from threadIdx, blockIdx, blockDim and gridDim.
//!
const unsigned kThread = vf::gpu::getNodeIndex();
-
+
//////////////////////////////////////////////////////////////////////////
- //! - Return for non-fluid nodes
+ // run for all indices in size_Mat and fluid nodes
if (kThread >= numberOfFluidNodes)
return;
-
////////////////////////////////////////////////////////////////////////////////
//! - Get the node index from the array containing all indices of fluid nodes
//!
@@ -85,10 +108,9 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
//! DOI:10.3390/computation5020019 ]</b></a>
//!
Distributions27 dist = vf::gpu::getDistributionReferences27(distributions, numberOfLBnodes, isEvenTimestep);
-
+
////////////////////////////////////////////////////////////////////////////////
//! - Set neighbor indices (necessary for indirect addressing)
- //!
uint k_M00 = neighborX[k_000];
uint k_0M0 = neighborY[k_000];
uint k_00M = neighborZ[k_000];
@@ -96,9 +118,8 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
uint k_M0M = neighborZ[k_M00];
uint k_0MM = neighborZ[k_0M0];
uint k_MMM = neighborZ[k_MM0];
-
////////////////////////////////////////////////////////////////////////////////////
- //! - Set local distributions (f's):
+ //! - Set local distributions
//!
real f_000 = (dist.f[DIR_000])[k_000];
real f_P00 = (dist.f[DIR_P00])[k_000];
@@ -159,45 +180,84 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
real& m_200 = f_PMM;
real& m_000 = f_MMM;
- ////////////////////////////////////////////////////////////////////////////////////
+ //////////////////////////////////////////////////////(unsigned long)//////////////////////////////
//! - Calculate density and velocity using pyramid summation for low round-off errors as in Eq. (J1)-(J3) \ref
//! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
//! DOI:10.1016/j.camwa.2015.05.001 ]</b></a>
//!
real drho = ((((f_PPP + f_MMM) + (f_MPM + f_PMP)) + ((f_MPP + f_PMM) + (f_MMP + f_PPM))) +
- (((f_0MP + f_0PM) + (f_0MM + f_0PP)) + ((f_M0P + f_P0M) + (f_M0M + f_P0P)) +
- ((f_MP0 + f_PM0) + (f_MM0 + f_PP0))) +
- ((f_M00 + f_P00) + (f_0M0 + f_0P0) + (f_00M + f_00P))) +
+ (((f_0MP + f_0PM) + (f_0MM + f_0PP)) + ((f_M0P + f_P0M) + (f_M0M + f_P0P)) +
+ ((f_MP0 + f_PM0) + (f_MM0 + f_PP0))) +
+ ((f_M00 + f_P00) + (f_0M0 + f_0P0) + (f_00M + f_00P))) +
f_000;
real oneOverRho = c1o1 / (c1o1 + drho);
real vvx = ((((f_PPP - f_MMM) + (f_PMP - f_MPM)) + ((f_PMM - f_MPP) + (f_PPM - f_MMP))) +
(((f_P0M - f_M0P) + (f_P0P - f_M0M)) + ((f_PM0 - f_MP0) + (f_PP0 - f_MM0))) + (f_P00 - f_M00)) *
- oneOverRho;
+ oneOverRho;
real vvy = ((((f_PPP - f_MMM) + (f_MPM - f_PMP)) + ((f_MPP - f_PMM) + (f_PPM - f_MMP))) +
(((f_0PM - f_0MP) + (f_0PP - f_0MM)) + ((f_MP0 - f_PM0) + (f_PP0 - f_MM0))) + (f_0P0 - f_0M0)) *
- oneOverRho;
+ oneOverRho;
real vvz = ((((f_PPP - f_MMM) + (f_PMP - f_MPM)) + ((f_MPP - f_PMM) + (f_MMP - f_PPM))) +
(((f_0MP - f_0PM) + (f_0PP - f_0MM)) + ((f_M0P - f_P0M) + (f_P0P - f_M0M))) + (f_00P - f_00M)) *
- oneOverRho;
-
+ oneOverRho;
+
////////////////////////////////////////////////////////////////////////////////////
//! - Add half of the acceleration (body force) to the velocity as in Eq. (42) \ref
//! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
//! DOI:10.1016/j.camwa.2015.05.001 ]</b></a>
//!
real factor = c1o1;
- // The factor has to be scaled for each level to get the correct acceleration.
for (size_t i = 1; i <= level; i++) {
factor *= c2o1;
}
- real fx = forces[0] / factor;
- real fy = forces[1] / factor;
- real fz = forces[2] / factor;
- vvx += fx * c1o2;
- vvy += fy * c1o2;
- vvz += fz * c1o2;
+
+ real fx = forces[0];
+ real fy = forces[1];
+ real fz = forces[2];
+
+ if( applyBodyForce ){
+ fx += bodyForceX[k_000];
+ fy += bodyForceY[k_000];
+ fz += bodyForceZ[k_000];
+
+ // real vx = vvx;
+ // real vy = vvy;
+ // real vz = vvz;
+ real acc_x = fx * c1o2 / factor;
+ real acc_y = fy * c1o2 / factor;
+ real acc_z = fz * c1o2 / factor;
+
+ vvx += acc_x;
+ vvy += acc_y;
+ vvz += acc_z;
+
+ // Reset body force. To be used when not using round-off correction.
+ bodyForceX[k_000] = 0.0f;
+ bodyForceY[k_000] = 0.0f;
+ bodyForceZ[k_000] = 0.0f;
+
+ ////////////////////////////////////////////////////////////////////////////////////
+ //!> Round-off correction
+ //!
+ //!> Similar to Kahan summation algorithm (https://en.wikipedia.org/wiki/Kahan_summation_algorithm)
+ //!> Essentially computes the round-off error of the applied force and adds it in the next time step as a compensation.
+ //!> Seems to be necesseary at very high Re boundary layers, where the forcing and velocity can
+ //!> differ by several orders of magnitude.
+ //!> \note 16/05/2022: Testing, still ongoing!
+ //!
+ // bodyForceX[k_000] = (acc_x-(vvx-vx))*factor*c2o1;
+ // bodyForceY[k_000] = (acc_y-(vvy-vy))*factor*c2o1;
+ // bodyForceZ[k_000] = (acc_z-(vvz-vz))*factor*c2o1;
+ }
+ else{
+ vvx += fx * c1o2 / factor;
+ vvy += fy * c1o2 / factor;
+ vvz += fz * c1o2 / factor;
+ }
+
+
////////////////////////////////////////////////////////////////////////////////////
// calculate the square of velocities for this lattice node
real vx2 = vvx * vvx;
@@ -272,15 +332,21 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
//! - Fifth order cumulants \f$ C_{221}, C_{212}, C_{122}\f$: \f$\omega_9=O5=1.0\f$.
//! - Sixth order cumulant \f$ C_{222}\f$: \f$\omega_{10}=O6=1.0\f$.
//!
+ ////////////////////////////////////////////////////////////////////////////////////
+ //! - Calculate modified omega with turbulent viscosity
+ //!
+ real omega = omega_in;
+ if(turbulenceModel != TurbulenceModel::None){ omega /= (c1o1 + c3o1*omega_in*turbulentViscosity[k_000]); }
////////////////////////////////////////////////////////////
// 2.
real OxxPyyPzz = c1o1;
////////////////////////////////////////////////////////////
// 3.
- real OxyyPxzz = c8o1 * (-c2o1 + omega) * (c1o1 + c2o1 * omega) / (-c8o1 - c14o1 * omega + c7o1 * omega * omega);
- real OxyyMxzz = c8o1 * (-c2o1 + omega) * (-c7o1 + c4o1 * omega) / (c56o1 - c50o1 * omega + c9o1 * omega * omega);
- real Oxyz = c24o1 * (-c2o1 + omega) * (-c2o1 - c7o1 * omega + c3o1 * omega * omega) /
- (c48o1 + c152o1 * omega - c130o1 * omega * omega + c29o1 * omega * omega * omega);
+ real OxyyPxzz = c8o1 * (-c2o1 + omega) * (c1o1 + c2o1 * omega) / (-c8o1 - c14o1 * omega + c7o1 * omega * omega);
+ real OxyyMxzz =
+ c8o1 * (-c2o1 + omega) * (-c7o1 + c4o1 * omega) / (c56o1 - c50o1 * omega + c9o1 * omega * omega);
+ real Oxyz = c24o1 * (-c2o1 + omega) * (-c2o1 - c7o1 * omega + c3o1 * omega * omega) /
+ (c48o1 + c152o1 * omega - c130o1 * omega * omega + c29o1 * omega * omega * omega);
////////////////////////////////////////////////////////////
// 4.
real O4 = c1o1;
@@ -292,16 +358,16 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
real O6 = c1o1;
////////////////////////////////////////////////////////////////////////////////////
- //! - A and B: parameters for fourth order convergence of the diffusion term according to Eq. (115) and (116)
+ //! - A and DIR_00M: parameters for fourth order convergence of the diffusion term according to Eq. (115) and (116)
//! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
//! DOI:10.1016/j.jcp.2017.05.040 ]</b></a> with simplifications assuming \f$ \omega_2 = 1.0 \f$ (modify for
//! different bulk viscosity).
//!
- real factorA = (c4o1 + c2o1 * omega - c3o1 * omega * omega) / (c2o1 - c7o1 * omega + c5o1 * omega * omega);
+ real factorA = (c4o1 + c2o1 * omega - c3o1 * omega * omega) / (c2o1 - c7o1 * omega + c5o1 * omega * omega);
real factorB = (c4o1 + c28o1 * omega - c14o1 * omega * omega) / (c6o1 - c21o1 * omega + c15o1 * omega * omega);
////////////////////////////////////////////////////////////////////////////////////
- //! - Compute cumulants (c's) from central moments according to Eq. (20)-(23) in
+ //! - Compute cumulants from central moments according to Eq. (20)-(23) in
//! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
//! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
//!
@@ -318,27 +384,27 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
// 5.
real c_122 =
m_122 - ((m_002 * m_120 + m_020 * m_102 + c4o1 * m_011 * m_111 + c2o1 * (m_101 * m_021 + m_110 * m_012)) +
- c1o3 * (m_120 + m_102)) *
- oneOverRho;
+ c1o3 * (m_120 + m_102)) *
+ oneOverRho;
real c_212 =
m_212 - ((m_002 * m_210 + m_200 * m_012 + c4o1 * m_101 * m_111 + c2o1 * (m_011 * m_201 + m_110 * m_102)) +
- c1o3 * (m_210 + m_012)) *
- oneOverRho;
+ c1o3 * (m_210 + m_012)) *
+ oneOverRho;
real c_221 =
m_221 - ((m_200 * m_021 + m_020 * m_201 + c4o1 * m_110 * m_111 + c2o1 * (m_101 * m_120 + m_011 * m_210)) +
- c1o3 * (m_021 + m_201)) *
- oneOverRho;
+ c1o3 * (m_021 + m_201)) *
+ oneOverRho;
////////////////////////////////////////////////////////////
// 6.
real c_222 = m_222 + ((-c4o1 * m_111 * m_111 - (m_200 * m_022 + m_020 * m_202 + m_002 * m_220) -
c4o1 * (m_011 * m_211 + m_101 * m_121 + m_110 * m_112) -
c2o1 * (m_120 * m_102 + m_210 * m_012 + m_201 * m_021)) *
oneOverRho +
- (c4o1 * (m_101 * m_101 * m_020 + m_011 * m_011 * m_200 + m_110 * m_110 * m_002) +
+ (c4o1 * (m_101 * m_101 * m_020 + m_011 * m_011 * m_200 + m_110 * m_110 * m_002) +
c2o1 * (m_200 * m_020 * m_002) + c16o1 * m_110 * m_101 * m_011) *
oneOverRho * oneOverRho -
c1o3 * (m_022 + m_202 + m_220) * oneOverRho - c1o9 * (m_200 + m_020 + m_002) * oneOverRho +
- (c2o1 * (m_101 * m_101 + m_011 * m_011 + m_110 * m_110) +
+ (c2o1 * (m_101 * m_101 + m_011 * m_011 + m_110 * m_110) +
(m_002 * m_020 + m_002 * m_200 + m_020 * m_200) + c1o3 * (m_002 + m_020 + m_200)) *
oneOverRho * oneOverRho * c2o3 +
c1o27 * ((drho * drho - drho) * oneOverRho * oneOverRho));
@@ -378,6 +444,22 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
real dxux = c1o2 * (-omega) * (mxxMyy + mxxMzz) + c1o2 * OxxPyyPzz * (m_000 - mxxPyyPzz);
real dyuy = dxux + omega * c3o2 * mxxMyy;
real dzuz = dxux + omega * c3o2 * mxxMzz;
+
+ ////////////////////////////////////////////////////////////////////////////////////
+ switch (turbulenceModel)
+ {
+ case TurbulenceModel::None:
+ case TurbulenceModel::AMD: //AMD is computed in separate kernel
+ break;
+ case TurbulenceModel::Smagorinsky:
+ turbulentViscosity[k_000] = calcTurbulentViscositySmagorinsky(SGSconstant, dxux, dyuy, dzuz, Dxy, Dxz , Dyz);
+ break;
+ case TurbulenceModel::QR:
+ turbulentViscosity[k_000] = calcTurbulentViscosityQR(SGSconstant, dxux, dyuy, dzuz, Dxy, Dxz , Dyz);
+ break;
+ default:
+ break;
+ }
////////////////////////////////////////////////////////////
//! - Relaxation of second order cumulants with correction terms according to Eq. (33)-(35) in
//! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
@@ -386,7 +468,6 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
mxxPyyPzz += OxxPyyPzz * (m_000 - mxxPyyPzz) - c3o1 * (c1o1 - c1o2 * OxxPyyPzz) * (vx2 * dxux + vy2 * dyuy + vz2 * dzuz);
mxxMyy += omega * (-mxxMyy) - c3o1 * (c1o1 + c1o2 * (-omega)) * (vx2 * dxux - vy2 * dyuy);
mxxMzz += omega * (-mxxMzz) - c3o1 * (c1o1 + c1o2 * (-omega)) * (vx2 * dxux - vz2 * dzuz);
- //////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////
////no correction
@@ -394,18 +475,18 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
// mxxMyy += -(-omega) * (-mxxMyy);
// mxxMzz += -(-omega) * (-mxxMzz);
//////////////////////////////////////////////////////////////////////////
-
m_011 += omega * (-m_011);
m_101 += omega * (-m_101);
m_110 += omega * (-m_110);
+ ////////////////////////////////////////////////////////////////////////////////////
+ // relax
//////////////////////////////////////////////////////////////////////////
+ // incl. limiter
//! - Relaxation of third order cumulants including limiter according to Eq. (116)-(123)
//! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
//! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
//!
- //////////////////////////////////////////////////////////////////////////
- // incl. limiter
real wadjust = Oxyz + (c1o1 - Oxyz) * abs(m_111) / (abs(m_111) + quadricLimitD);
m_111 += wadjust * (-m_111);
wadjust = OxyyPxzz + (c1o1 - OxyyPxzz) * abs(mxxyPyzz) / (abs(mxxyPyzz) + quadricLimitP);
@@ -459,6 +540,7 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
c_121 = -O4 * (c1o1 / omega - c1o2) * Dxz * c1o3 * factorB + (c1o1 - O4) * (c_121);
c_211 = -O4 * (c1o1 / omega - c1o2) * Dyz * c1o3 * factorB + (c1o1 - O4) * (c_211);
+
//////////////////////////////////////////////////////////////////////////
// 5.
c_122 += O5 * (-c_122);
@@ -503,17 +585,17 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
//////////////////////////////////////////////////////////////////////////
// 6.
m_222 = c_222 - ((-c4o1 * m_111 * m_111 - (m_200 * m_022 + m_020 * m_202 + m_002 * m_220) -
- c4o1 * (m_011 * m_211 + m_101 * m_121 + m_110 * m_112) -
- c2o1 * (m_120 * m_102 + m_210 * m_012 + m_201 * m_021)) *
- oneOverRho +
- (c4o1 * (m_101 * m_101 * m_020 + m_011 * m_011 * m_200 + m_110 * m_110 * m_002) +
- c2o1 * (m_200 * m_020 * m_002) + c16o1 * m_110 * m_101 * m_011) *
- oneOverRho * oneOverRho -
- c1o3 * (m_022 + m_202 + m_220) * oneOverRho - c1o9 * (m_200 + m_020 + m_002) * oneOverRho +
- (c2o1 * (m_101 * m_101 + m_011 * m_011 + m_110 * m_110) +
- (m_002 * m_020 + m_002 * m_200 + m_020 * m_200) + c1o3 * (m_002 + m_020 + m_200)) *
- oneOverRho * oneOverRho * c2o3 +
- c1o27 * ((drho * drho - drho) * oneOverRho * oneOverRho));
+ c4o1 * (m_011 * m_211 + m_101 * m_121 + m_110 * m_112) -
+ c2o1 * (m_120 * m_102 + m_210 * m_012 + m_201 * m_021)) *
+ oneOverRho +
+ (c4o1 * (m_101 * m_101 * m_020 + m_011 * m_011 * m_200 + m_110 * m_110 * m_002) +
+ c2o1 * (m_200 * m_020 * m_002) + c16o1 * m_110 * m_101 * m_011) *
+ oneOverRho * oneOverRho -
+ c1o3 * (m_022 + m_202 + m_220) * oneOverRho - c1o9 * (m_200 + m_020 + m_002) * oneOverRho +
+ (c2o1 * (m_101 * m_101 + m_011 * m_011 + m_110 * m_110) +
+ (m_002 * m_020 + m_002 * m_200 + m_020 * m_200) + c1o3 * (m_002 + m_020 + m_200)) *
+ oneOverRho * oneOverRho * c2o3 +
+ c1o27 * ((drho * drho - drho) * oneOverRho * oneOverRho));
////////////////////////////////////////////////////////////////////////////////////
//! - Add acceleration (body force) to first order cumulants according to Eq. (85)-(87) in
@@ -524,6 +606,15 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
m_010 = -m_010;
m_001 = -m_001;
+ //Write to array here to distribute read/write
+ if(writeMacroscopicVariables)
+ {
+ rho[k_000] = drho;
+ vx[k_000] = vvx;
+ vy[k_000] = vvy;
+ vz[k_000] = vvz;
+ }
+
////////////////////////////////////////////////////////////////////////////////////
//! - Chimera transform from central moments to well conditioned distributions as defined in Appendix J in
//! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
@@ -573,31 +664,63 @@ __global__ void LB_Kernel_CumulantK17CompChimRedesigned(
//! <a href="https://doi.org/10.3390/computation5020019"><b>[ M. Geier et al. (2017),
//! DOI:10.3390/computation5020019 ]</b></a>
//!
- (dist.f[DIR_P00])[k_000] = f_M00;
- (dist.f[DIR_M00])[k_M00] = f_P00;
- (dist.f[DIR_0P0])[k_000] = f_0M0;
- (dist.f[DIR_0M0])[k_0M0] = f_0P0;
- (dist.f[DIR_00P])[k_000] = f_00M;
- (dist.f[DIR_00M])[k_00M] = f_00P;
- (dist.f[DIR_PP0])[k_000] = f_MM0;
- (dist.f[DIR_MM0])[k_MM0] = f_PP0;
- (dist.f[DIR_PM0])[k_0M0] = f_MP0;
- (dist.f[DIR_MP0])[k_M00] = f_PM0;
- (dist.f[DIR_P0P])[k_000] = f_M0M;
- (dist.f[DIR_M0M])[k_M0M] = f_P0P;
- (dist.f[DIR_P0M])[k_00M] = f_M0P;
- (dist.f[DIR_M0P])[k_M00] = f_P0M;
- (dist.f[DIR_0PP])[k_000] = f_0MM;
- (dist.f[DIR_0MM])[k_0MM] = f_0PP;
- (dist.f[DIR_0PM])[k_00M] = f_0MP;
- (dist.f[DIR_0MP])[k_0M0] = f_0PM;
+ (dist.f[DIR_P00])[k_000] = f_M00;
+ (dist.f[DIR_M00])[k_M00] = f_P00;
+ (dist.f[DIR_0P0])[k_000] = f_0M0;
+ (dist.f[DIR_0M0])[k_0M0] = f_0P0;
+ (dist.f[DIR_00P])[k_000] = f_00M;
+ (dist.f[DIR_00M])[k_00M] = f_00P;
+ (dist.f[DIR_PP0])[k_000] = f_MM0;
+ (dist.f[DIR_MM0])[k_MM0] = f_PP0;
+ (dist.f[DIR_PM0])[k_0M0] = f_MP0;
+ (dist.f[DIR_MP0])[k_M00] = f_PM0;
+ (dist.f[DIR_P0P])[k_000] = f_M0M;
+ (dist.f[DIR_M0M])[k_M0M] = f_P0P;
+ (dist.f[DIR_P0M])[k_00M] = f_M0P;
+ (dist.f[DIR_M0P])[k_M00] = f_P0M;
+ (dist.f[DIR_0PP])[k_000] = f_0MM;
+ (dist.f[DIR_0MM])[k_0MM] = f_0PP;
+ (dist.f[DIR_0PM])[k_00M] = f_0MP;
+ (dist.f[DIR_0MP])[k_0M0] = f_0PM;
(dist.f[DIR_000])[k_000] = f_000;
- (dist.f[DIR_PPP])[k_000] = f_MMM;
- (dist.f[DIR_PMP])[k_0M0] = f_MPM;
- (dist.f[DIR_PPM])[k_00M] = f_MMP;
- (dist.f[DIR_PMM])[k_0MM] = f_MPP;
- (dist.f[DIR_MPP])[k_M00] = f_PMM;
- (dist.f[DIR_MMP])[k_MM0] = f_PPM;
- (dist.f[DIR_MPM])[k_M0M] = f_PMP;
- (dist.f[DIR_MMM])[k_MMM] = f_PPP;
-}
\ No newline at end of file
+ (dist.f[DIR_PPP])[k_000] = f_MMM;
+ (dist.f[DIR_PMP])[k_0M0] = f_MPM;
+ (dist.f[DIR_PPM])[k_00M] = f_MMP;
+ (dist.f[DIR_PMM])[k_0MM] = f_MPP;
+ (dist.f[DIR_MPP])[k_M00] = f_PMM;
+ (dist.f[DIR_MMP])[k_MM0] = f_PPM;
+ (dist.f[DIR_MPM])[k_M0M] = f_PMP;
+ (dist.f[DIR_MMM])[k_MMM] = f_PPP;
+}
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::AMD, true, true > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::Smagorinsky, true, true > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::QR, true, true > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::None, true, true > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::AMD, true, false > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::Smagorinsky, true, false > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::QR, true, false > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::None, true, false > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::AMD, false, true > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::Smagorinsky, false, true > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::QR, false, true > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::None, false, true > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::AMD, false, false > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::Smagorinsky, false, false > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::QR, false, false > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
+
+template __global__ void LB_Kernel_CumulantK17 < TurbulenceModel::None, false, false > ( real omega_in, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep, const uint *fluidNodeIndices, uint numberOfFluidNodes);
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim_Device.cuh b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17_Device.cuh
similarity index 51%
rename from src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim_Device.cuh
rename to src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17_Device.cuh
index 5ef37557399f263d25edf03b02b00f6a03c6e1cb..b8cc9543e9b531c5aa90cb2961416a6cbc52377d 100644
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim_Device.cuh
+++ b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17_Device.cuh
@@ -1,12 +1,11 @@
-#ifndef LB_Kernel_TURBULENT_VISCOSITY_CUMULANT_K17_COMP_CHIM_H
-#define LB_Kernel_TURBULENT_VISCOSITY_CUMULANT_K17_COMP_CHIM_H
+#ifndef LB_Kernel_CUMULANT_K17_H
+#define LB_Kernel_CUMULANT_K17_H
#include <DataTypes.h>
#include <curand.h>
-template< TurbulenceModel turbulenceModel > __global__ void LB_Kernel_TurbulentViscosityCumulantK17CompChim(
+template< TurbulenceModel turbulenceModel, bool writeMacroscopicVariables, bool applyBodyForce > __global__ void LB_Kernel_CumulantK17(
real omega_in,
- uint* typeOfGridNode,
uint* neighborX,
uint* neighborY,
uint* neighborZ,
@@ -17,13 +16,14 @@ template< TurbulenceModel turbulenceModel > __global__ void LB_Kernel_TurbulentV
real* vz,
real* turbulentViscosity,
real SGSconstant,
- unsigned long size_Mat,
+ unsigned long numberOfLBnodes,
int level,
- bool bodyForce,
real* forces,
real* bodyForceX,
real* bodyForceY,
real* bodyForceZ,
real* quadricLimiters,
- bool isEvenTimestep);
+ bool isEvenTimestep,
+ const uint *fluidNodeIndices,
+ uint numberOfFluidNodes);
#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned.cu
deleted file mode 100644
index 8c06b7117c8b1ef62b932a76bf5de0be2ae99b1c..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned.cu
+++ /dev/null
@@ -1,61 +0,0 @@
-#include "CumulantK17CompChimRedesigned.h"
-
-#include "Parameter/Parameter.h"
-#include "Parameter/CudaStreamManager.h"
-#include "CumulantK17CompChimRedesigned_Device.cuh"
-
-#include <cuda.h>
-
-std::shared_ptr<CumulantK17CompChimRedesigned> CumulantK17CompChimRedesigned::getNewInstance(std::shared_ptr<Parameter> para,
- int level)
-{
- return std::shared_ptr<CumulantK17CompChimRedesigned>(new CumulantK17CompChimRedesigned(para, level));
-}
-
-void CumulantK17CompChimRedesigned::run()
-{
- LB_Kernel_CumulantK17CompChimRedesigned <<< cudaGrid.grid, cudaGrid.threads >>>(
- para->getParD(level)->omega,
- para->getParD(level)->neighborX,
- para->getParD(level)->neighborY,
- para->getParD(level)->neighborZ,
- para->getParD(level)->distributions.f[0],
- para->getParD(level)->numberOfNodes,
- level,
- para->getForcesDev(),
- para->getQuadricLimitersDev(),
- para->getParD(level)->isEvenTimestep,
- para->getParD(level)->fluidNodeIndices,
- para->getParD(level)->numberOfFluidNodes);
- getLastCudaError("LB_Kernel_CumulantK17CompChim execution failed");
-}
-
-void CumulantK17CompChimRedesigned::runOnIndices(const unsigned int *indices, unsigned int size_indices, int streamIndex)
-{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
-
- LB_Kernel_CumulantK17CompChimRedesigned<<< cudaGrid.grid, cudaGrid.threads, 0, stream>>>(
- para->getParD(level)->omega,
- para->getParD(level)->neighborX,
- para->getParD(level)->neighborY,
- para->getParD(level)->neighborZ,
- para->getParD(level)->distributions.f[0],
- para->getParD(level)->numberOfNodes,
- level,
- para->getForcesDev(),
- para->getQuadricLimitersDev(),
- para->getParD(level)->isEvenTimestep,
- indices,
- size_indices);
- getLastCudaError("LB_Kernel_CumulantK17CompChim execution failed");
-
-}
-
-CumulantK17CompChimRedesigned::CumulantK17CompChimRedesigned(std::shared_ptr<Parameter> para, int level): KernelImp(para, level)
-{
- myPreProcessorTypes.push_back(InitCompSP27);
- myKernelGroup = BasicKernel;
- this->cudaGrid = vf::cuda::CudaGrid(para->getParD(level)->numberofthreads, para->getParD(level)->numberOfNodes);
- this->kernelUsesFluidNodeIndices = true;
-}
-
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned.h b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned.h
deleted file mode 100644
index 4658075de330665fdba88a5ec8149a9b476d5ac7..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned.h
+++ /dev/null
@@ -1,18 +0,0 @@
-#ifndef CUMULANT_K17_COMP_CHIM_REDESIGN_H
-#define CUMULANT_K17_COMP_CHIM_REDESIGN_H
-
-#include "Kernel/KernelImp.h"
-
-class CumulantK17CompChimRedesigned : public KernelImp
-{
-public:
- static std::shared_ptr<CumulantK17CompChimRedesigned> getNewInstance(std::shared_ptr<Parameter> para, int level);
- void run() override;
- void runOnIndices(const unsigned int *indices, unsigned int size_indices, int stream = -1) override;
-
-private:
- CumulantK17CompChimRedesigned();
- CumulantK17CompChimRedesigned(std::shared_ptr<Parameter> para, int level);
-};
-
-#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned_Device.cuh b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned_Device.cuh
deleted file mode 100644
index 00628efc76447a09504d2fd32a26a63a4d611c66..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned_Device.cuh
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef LB_Kernel_CUMULANT_K17_COMP_CHIM_REDESIGN_H
-#define LB_Kernel_CUMULANT_K17_COMP_CHIM_REDESIGN_H
-
-#include <DataTypes.h>
-#include <curand.h>
-
-__global__ void LB_Kernel_CumulantK17CompChimRedesigned(
- real omega,
- uint* neighborX,
- uint* neighborY,
- uint* neighborZ,
- real* distributions,
- unsigned long numberOfLBnodes,
- int level,
- real* forces,
- real* quadricLimiters,
- bool isEvenTimestep,
- const uint* fluidNodeIndices,
- uint numberOfFluidNodes);
-#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream.cu
deleted file mode 100644
index 6fae9f6d4845019afd363790eea0ee17c69a060f..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream.cu
+++ /dev/null
@@ -1,61 +0,0 @@
-#include "CumulantK17CompChimStream.h"
-
-#include "Parameter/Parameter.h"
-#include "Parameter/CudaStreamManager.h"
-#include "CumulantK17CompChimStream_Device.cuh"
-
-#include <cuda.h>
-
-std::shared_ptr<CumulantK17CompChimStream> CumulantK17CompChimStream::getNewInstance(std::shared_ptr<Parameter> para,
- int level)
-{
- return std::shared_ptr<CumulantK17CompChimStream>(new CumulantK17CompChimStream(para, level));
-}
-
-void CumulantK17CompChimStream::run()
-{
- LB_Kernel_CumulantK17CompChimStream <<< cudaGrid.grid, cudaGrid.threads >>>(
- para->getParD(level)->omega,
- para->getParD(level)->neighborX,
- para->getParD(level)->neighborY,
- para->getParD(level)->neighborZ,
- para->getParD(level)->distributions.f[0],
- para->getParD(level)->numberOfNodes,
- level,
- para->getForcesDev(),
- para->getQuadricLimitersDev(),
- para->getParD(level)->isEvenTimestep,
- para->getParD(level)->fluidNodeIndices,
- para->getParD(level)->numberOfFluidNodes);
- getLastCudaError("LB_Kernel_CumulantK17CompChim execution failed");
-}
-
-void CumulantK17CompChimStream::runOnIndices(const unsigned int *indices, unsigned int size_indices, int streamIndex)
-{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
-
- LB_Kernel_CumulantK17CompChimStream<<< cudaGrid.grid, cudaGrid.threads, 0, stream>>>(
- para->getParD(level)->omega,
- para->getParD(level)->neighborX,
- para->getParD(level)->neighborY,
- para->getParD(level)->neighborZ,
- para->getParD(level)->distributions.f[0],
- para->getParD(level)->numberOfNodes,
- level,
- para->getForcesDev(),
- para->getQuadricLimitersDev(),
- para->getParD(level)->isEvenTimestep,
- indices,
- size_indices);
- getLastCudaError("LB_Kernel_CumulantK17CompChim execution failed");
-
-}
-
-CumulantK17CompChimStream::CumulantK17CompChimStream(std::shared_ptr<Parameter> para, int level): KernelImp(para, level)
-{
- myPreProcessorTypes.push_back(InitCompSP27);
- myKernelGroup = BasicKernel;
- this->cudaGrid = vf::cuda::CudaGrid(para->getParD(level)->numberofthreads, para->getParD(level)->numberOfNodes);
- this->kernelUsesFluidNodeIndices = true;
-}
-
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream.h b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream.h
deleted file mode 100644
index 325826e04c893b7c56b7f00bb2503a4eb1fda441..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream.h
+++ /dev/null
@@ -1,18 +0,0 @@
-#ifndef CUMULANT_K17_COMP_CHIM_SPARSE_H
-#define CUMULANT_K17_COMP_CHIM_SPARSE_H
-
-#include "Kernel/KernelImp.h"
-
-class CumulantK17CompChimStream : public KernelImp
-{
-public:
- static std::shared_ptr<CumulantK17CompChimStream> getNewInstance(std::shared_ptr<Parameter> para, int level);
- void run() override;
- void runOnIndices(const unsigned int *indices, unsigned int size_indices, int stream = -1) override;
-
-private:
- CumulantK17CompChimStream();
- CumulantK17CompChimStream(std::shared_ptr<Parameter> para, int level);
-};
-
-#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStreamDevice.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStreamDevice.cu
deleted file mode 100644
index 830fcc6c328f2ecd0f626539040868696065065f..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStreamDevice.cu
+++ /dev/null
@@ -1,640 +0,0 @@
-//=======================================================================================
-// ____ ____ __ ______ __________ __ __ __ __
-// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
-// \ \ | | | | | |_) | | | | | | | / \ | |
-// \ \ | | | | | _ / | | | | | | / /\ \ | |
-// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
-// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
-// \ \ | | ________________________________________________________________
-// \ \ | | | ______________________________________________________________|
-// \ \| | | | __ __ __ __ ______ _______
-// \ | | |_____ | | | | | | | | | _ \ / _____)
-// \ | | _____| | | | | | | | | | | \ \ \_______
-// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
-// \ _____| |__| |________| \_______/ |__| |______/ (_______/
-//
-// This file is part of VirtualFluids. VirtualFluids is free software: you can
-// redistribute it and/or modify it under the terms of the GNU General Public
-// License as published by the Free Software Foundation, either version 3 of
-// the License, or (at your option) any later version.
-//
-// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
-// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
-// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
-// for more details.
-//
-// You should have received a copy of the GNU General Public License along
-// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
-//
-//! \file Cumulant27chimStream.cu
-//! \ingroup GPU
-//! \author Martin Schoenherr, Anna Wellmann
-//=======================================================================================
-/* Device code */
-#include "LBM/LB.h"
-#include "lbm/constants/D3Q27.h"
-#include <lbm/constants/NumericConstants.h>
-
-using namespace vf::lbm::constant;
-using namespace vf::lbm::dir;
-#include "Kernel/Utilities/ChimeraTransformation.h"
-
-////////////////////////////////////////////////////////////////////////////////
-__global__ void LB_Kernel_CumulantK17CompChimStream(
- real omega,
- uint* neighborX,
- uint* neighborY,
- uint* neighborZ,
- real* distributions,
- unsigned long size_Mat,
- int level,
- real* forces,
- real* quadricLimiters,
- bool isEvenTimestep,
- const uint *fluidNodeIndices,
- uint numberOfFluidNodes)
-{
- //////////////////////////////////////////////////////////////////////////
- //! Cumulant K17 Kernel is based on \ref
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017), DOI:10.1016/j.jcp.2017.05.040
- //! ]</b></a> and \ref <a href="https://doi.org/10.1016/j.jcp.2017.07.004"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.07.004 ]</b></a>
- //!
- //! The cumulant kernel is executed in the following steps
- //!
- ////////////////////////////////////////////////////////////////////////////////
- //! - Get node index coordinates from threadIdx, blockIdx, blockDim and gridDim.
- //!
- const unsigned x = threadIdx.x;
- const unsigned y = blockIdx.x;
- const unsigned z = blockIdx.y;
-
- const unsigned nx = blockDim.x;
- const unsigned ny = gridDim.x;
-
- const unsigned k_thread = nx * (ny * z + y) + x;
-
- //////////////////////////////////////////////////////////////////////////
- // run for all indices in fluidNodeIndices
- if (k_thread < numberOfFluidNodes) {
- //////////////////////////////////////////////////////////////////////////
- //! - Read distributions: style of reading and writing the distributions from/to stored arrays dependent on
- //! timestep is based on the esoteric twist algorithm \ref <a
- //! href="https://doi.org/10.3390/computation5020019"><b>[ M. Geier et al. (2017),
- //! DOI:10.3390/computation5020019 ]</b></a>
-
- const unsigned k = fluidNodeIndices[k_thread];
-
- Distributions27 dist;
- if (isEvenTimestep) {
- dist.f[DIR_P00] = &distributions[DIR_P00 * size_Mat];
- dist.f[DIR_M00] = &distributions[DIR_M00 * size_Mat];
- dist.f[DIR_0P0] = &distributions[DIR_0P0 * size_Mat];
- dist.f[DIR_0M0] = &distributions[DIR_0M0 * size_Mat];
- dist.f[DIR_00P] = &distributions[DIR_00P * size_Mat];
- dist.f[DIR_00M] = &distributions[DIR_00M * size_Mat];
- dist.f[DIR_PP0] = &distributions[DIR_PP0 * size_Mat];
- dist.f[DIR_MM0] = &distributions[DIR_MM0 * size_Mat];
- dist.f[DIR_PM0] = &distributions[DIR_PM0 * size_Mat];
- dist.f[DIR_MP0] = &distributions[DIR_MP0 * size_Mat];
- dist.f[DIR_P0P] = &distributions[DIR_P0P * size_Mat];
- dist.f[DIR_M0M] = &distributions[DIR_M0M * size_Mat];
- dist.f[DIR_P0M] = &distributions[DIR_P0M * size_Mat];
- dist.f[DIR_M0P] = &distributions[DIR_M0P * size_Mat];
- dist.f[DIR_0PP] = &distributions[DIR_0PP * size_Mat];
- dist.f[DIR_0MM] = &distributions[DIR_0MM * size_Mat];
- dist.f[DIR_0PM] = &distributions[DIR_0PM * size_Mat];
- dist.f[DIR_0MP] = &distributions[DIR_0MP * size_Mat];
- dist.f[DIR_000] = &distributions[DIR_000 * size_Mat];
- dist.f[DIR_PPP] = &distributions[DIR_PPP * size_Mat];
- dist.f[DIR_MMP] = &distributions[DIR_MMP * size_Mat];
- dist.f[DIR_PMP] = &distributions[DIR_PMP * size_Mat];
- dist.f[DIR_MPP] = &distributions[DIR_MPP * size_Mat];
- dist.f[DIR_PPM] = &distributions[DIR_PPM * size_Mat];
- dist.f[DIR_MMM] = &distributions[DIR_MMM * size_Mat];
- dist.f[DIR_PMM] = &distributions[DIR_PMM * size_Mat];
- dist.f[DIR_MPM] = &distributions[DIR_MPM * size_Mat];
- } else {
- dist.f[DIR_M00] = &distributions[DIR_P00 * size_Mat];
- dist.f[DIR_P00] = &distributions[DIR_M00 * size_Mat];
- dist.f[DIR_0M0] = &distributions[DIR_0P0 * size_Mat];
- dist.f[DIR_0P0] = &distributions[DIR_0M0 * size_Mat];
- dist.f[DIR_00M] = &distributions[DIR_00P * size_Mat];
- dist.f[DIR_00P] = &distributions[DIR_00M * size_Mat];
- dist.f[DIR_MM0] = &distributions[DIR_PP0 * size_Mat];
- dist.f[DIR_PP0] = &distributions[DIR_MM0 * size_Mat];
- dist.f[DIR_MP0] = &distributions[DIR_PM0 * size_Mat];
- dist.f[DIR_PM0] = &distributions[DIR_MP0 * size_Mat];
- dist.f[DIR_M0M] = &distributions[DIR_P0P * size_Mat];
- dist.f[DIR_P0P] = &distributions[DIR_M0M * size_Mat];
- dist.f[DIR_M0P] = &distributions[DIR_P0M * size_Mat];
- dist.f[DIR_P0M] = &distributions[DIR_M0P * size_Mat];
- dist.f[DIR_0MM] = &distributions[DIR_0PP * size_Mat];
- dist.f[DIR_0PP] = &distributions[DIR_0MM * size_Mat];
- dist.f[DIR_0MP] = &distributions[DIR_0PM * size_Mat];
- dist.f[DIR_0PM] = &distributions[DIR_0MP * size_Mat];
- dist.f[DIR_000] = &distributions[DIR_000 * size_Mat];
- dist.f[DIR_MMM] = &distributions[DIR_PPP * size_Mat];
- dist.f[DIR_PPM] = &distributions[DIR_MMP * size_Mat];
- dist.f[DIR_MPM] = &distributions[DIR_PMP * size_Mat];
- dist.f[DIR_PMM] = &distributions[DIR_MPP * size_Mat];
- dist.f[DIR_MMP] = &distributions[DIR_PPM * size_Mat];
- dist.f[DIR_PPP] = &distributions[DIR_MMM * size_Mat];
- dist.f[DIR_MPP] = &distributions[DIR_PMM * size_Mat];
- dist.f[DIR_PMP] = &distributions[DIR_MPM * size_Mat];
- }
- ////////////////////////////////////////////////////////////////////////////////
- //! - Set neighbor indices (necessary for indirect addressing)
- uint kw = neighborX[k];
- uint ks = neighborY[k];
- uint kb = neighborZ[k];
- uint ksw = neighborY[kw];
- uint kbw = neighborZ[kw];
- uint kbs = neighborZ[ks];
- uint kbsw = neighborZ[ksw];
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Set local distributions
- //!
- real mfcbb = (dist.f[DIR_P00])[k];
- real mfabb = (dist.f[DIR_M00])[kw];
- real mfbcb = (dist.f[DIR_0P0])[k];
- real mfbab = (dist.f[DIR_0M0])[ks];
- real mfbbc = (dist.f[DIR_00P])[k];
- real mfbba = (dist.f[DIR_00M])[kb];
- real mfccb = (dist.f[DIR_PP0])[k];
- real mfaab = (dist.f[DIR_MM0])[ksw];
- real mfcab = (dist.f[DIR_PM0])[ks];
- real mfacb = (dist.f[DIR_MP0])[kw];
- real mfcbc = (dist.f[DIR_P0P])[k];
- real mfaba = (dist.f[DIR_M0M])[kbw];
- real mfcba = (dist.f[DIR_P0M])[kb];
- real mfabc = (dist.f[DIR_M0P])[kw];
- real mfbcc = (dist.f[DIR_0PP])[k];
- real mfbaa = (dist.f[DIR_0MM])[kbs];
- real mfbca = (dist.f[DIR_0PM])[kb];
- real mfbac = (dist.f[DIR_0MP])[ks];
- real mfbbb = (dist.f[DIR_000])[k];
- real mfccc = (dist.f[DIR_PPP])[k];
- real mfaac = (dist.f[DIR_MMP])[ksw];
- real mfcac = (dist.f[DIR_PMP])[ks];
- real mfacc = (dist.f[DIR_MPP])[kw];
- real mfcca = (dist.f[DIR_PPM])[kb];
- real mfaaa = (dist.f[DIR_MMM])[kbsw];
- real mfcaa = (dist.f[DIR_PMM])[kbs];
- real mfaca = (dist.f[DIR_MPM])[kbw];
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Calculate density and velocity using pyramid summation for low round-off errors as in Eq. (J1)-(J3) \ref
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a>
- //!
- real drho = ((((mfccc + mfaaa) + (mfaca + mfcac)) + ((mfacc + mfcaa) + (mfaac + mfcca))) +
- (((mfbac + mfbca) + (mfbaa + mfbcc)) + ((mfabc + mfcba) + (mfaba + mfcbc)) +
- ((mfacb + mfcab) + (mfaab + mfccb))) +
- ((mfabb + mfcbb) + (mfbab + mfbcb) + (mfbba + mfbbc))) +
- mfbbb;
-
- real rho = c1o1 + drho;
- real OOrho = c1o1 / rho;
-
- real vvx = ((((mfccc - mfaaa) + (mfcac - mfaca)) + ((mfcaa - mfacc) + (mfcca - mfaac))) +
- (((mfcba - mfabc) + (mfcbc - mfaba)) + ((mfcab - mfacb) + (mfccb - mfaab))) + (mfcbb - mfabb)) *
- OOrho;
- real vvy = ((((mfccc - mfaaa) + (mfaca - mfcac)) + ((mfacc - mfcaa) + (mfcca - mfaac))) +
- (((mfbca - mfbac) + (mfbcc - mfbaa)) + ((mfacb - mfcab) + (mfccb - mfaab))) + (mfbcb - mfbab)) *
- OOrho;
- real vvz = ((((mfccc - mfaaa) + (mfcac - mfaca)) + ((mfacc - mfcaa) + (mfaac - mfcca))) +
- (((mfbac - mfbca) + (mfbcc - mfbaa)) + ((mfabc - mfcba) + (mfcbc - mfaba))) + (mfbbc - mfbba)) *
- OOrho;
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Add half of the acceleration (body force) to the velocity as in Eq. (42) \ref
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a>
- //!
- real factor = c1o1;
- for (size_t i = 1; i <= level; i++) {
- factor *= c2o1;
- }
- real fx = forces[0] / factor;
- real fy = forces[1] / factor;
- real fz = forces[2] / factor;
- vvx += fx * c1o2;
- vvy += fy * c1o2;
- vvz += fz * c1o2;
- ////////////////////////////////////////////////////////////////////////////////////
- // calculate the square of velocities for this lattice node
- real vx2 = vvx * vvx;
- real vy2 = vvy * vvy;
- real vz2 = vvz * vvz;
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Set relaxation limiters for third order cumulants to default value \f$ \lambda=0.001 \f$ according to
- //! section 6 in \ref <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- real wadjust;
- real qudricLimitP = quadricLimiters[0];
- real qudricLimitM = quadricLimiters[1];
- real qudricLimitD = quadricLimiters[2];
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Chimera transform from well conditioned distributions to central moments as defined in Appendix J in \ref
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a> see also Eq. (6)-(14) in \ref <a
- //! href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017), DOI:10.1016/j.jcp.2017.05.040
- //! ]</b></a>
- //!
- ////////////////////////////////////////////////////////////////////////////////////
- // Z - Dir
- forwardInverseChimeraWithK(mfaaa, mfaab, mfaac, vvz, vz2, c36o1, c1o36);
- forwardInverseChimeraWithK(mfaba, mfabb, mfabc, vvz, vz2, c9o1, c1o9);
- forwardInverseChimeraWithK(mfaca, mfacb, mfacc, vvz, vz2, c36o1, c1o36);
- forwardInverseChimeraWithK(mfbaa, mfbab, mfbac, vvz, vz2, c9o1, c1o9);
- forwardInverseChimeraWithK(mfbba, mfbbb, mfbbc, vvz, vz2, c9o4, c4o9);
- forwardInverseChimeraWithK(mfbca, mfbcb, mfbcc, vvz, vz2, c9o1, c1o9);
- forwardInverseChimeraWithK(mfcaa, mfcab, mfcac, vvz, vz2, c36o1, c1o36);
- forwardInverseChimeraWithK(mfcba, mfcbb, mfcbc, vvz, vz2, c9o1, c1o9);
- forwardInverseChimeraWithK(mfcca, mfccb, mfccc, vvz, vz2, c36o1, c1o36);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // Y - Dir
- forwardInverseChimeraWithK(mfaaa, mfaba, mfaca, vvy, vy2, c6o1, c1o6);
- forwardChimera(mfaab, mfabb, mfacb, vvy, vy2);
- forwardInverseChimeraWithK(mfaac, mfabc, mfacc, vvy, vy2, c18o1, c1o18);
- forwardInverseChimeraWithK(mfbaa, mfbba, mfbca, vvy, vy2, c3o2, c2o3);
- forwardChimera(mfbab, mfbbb, mfbcb, vvy, vy2);
- forwardInverseChimeraWithK(mfbac, mfbbc, mfbcc, vvy, vy2, c9o2, c2o9);
- forwardInverseChimeraWithK(mfcaa, mfcba, mfcca, vvy, vy2, c6o1, c1o6);
- forwardChimera(mfcab, mfcbb, mfccb, vvy, vy2);
- forwardInverseChimeraWithK(mfcac, mfcbc, mfccc, vvy, vy2, c18o1, c1o18);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // X - Dir
- forwardInverseChimeraWithK(mfaaa, mfbaa, mfcaa, vvx, vx2, c1o1, c1o1);
- forwardChimera(mfaba, mfbba, mfcba, vvx, vx2);
- forwardInverseChimeraWithK(mfaca, mfbca, mfcca, vvx, vx2, c3o1, c1o3);
- forwardChimera(mfaab, mfbab, mfcab, vvx, vx2);
- forwardChimera(mfabb, mfbbb, mfcbb, vvx, vx2);
- forwardChimera(mfacb, mfbcb, mfccb, vvx, vx2);
- forwardInverseChimeraWithK(mfaac, mfbac, mfcac, vvx, vx2, c3o1, c1o3);
- forwardChimera(mfabc, mfbbc, mfcbc, vvx, vx2);
- forwardInverseChimeraWithK(mfacc, mfbcc, mfccc, vvx, vx2, c3o1, c1o9);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Setting relaxation rates for non-hydrodynamic cumulants (default values). Variable names and equations
- //! according to <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //! => [NAME IN PAPER]=[NAME IN CODE]=[DEFAULT VALUE].
- //! - Trace of second order cumulants \f$ C_{200}+C_{020}+C_{002} \f$ used to adjust bulk
- //! viscosity:\f$\omega_2=OxxPyyPzz=1.0 \f$.
- //! - Third order cumulants \f$ C_{120}+C_{102}, C_{210}+C_{012}, C_{201}+C_{021} \f$: \f$ \omega_3=OxyyPxzz
- //! \f$ set according to Eq. (111) with simplifications assuming \f$ \omega_2=1.0\f$.
- //! - Third order cumulants \f$ C_{120}-C_{102}, C_{210}-C_{012}, C_{201}-C_{021} \f$: \f$ \omega_4 = OxyyMxzz
- //! \f$ set according to Eq. (112) with simplifications assuming \f$ \omega_2 = 1.0\f$.
- //! - Third order cumulants \f$ C_{111} \f$: \f$ \omega_5 = Oxyz \f$ set according to Eq. (113) with
- //! simplifications assuming \f$ \omega_2 = 1.0\f$ (modify for different bulk viscosity).
- //! - Fourth order cumulants \f$ C_{220}, C_{202}, C_{022}, C_{211}, C_{121}, C_{112} \f$: for simplification
- //! all set to the same default value \f$ \omega_6=\omega_7=\omega_8=O4=1.0 \f$.
- //! - Fifth order cumulants \f$ C_{221}, C_{212}, C_{122}\f$: \f$\omega_9=O5=1.0\f$.
- //! - Sixth order cumulant \f$ C_{222}\f$: \f$\omega_{10}=O6=1.0\f$.
- //!
- ////////////////////////////////////////////////////////////
- // 2.
- real OxxPyyPzz = c1o1;
- ////////////////////////////////////////////////////////////
- // 3.
- real OxyyPxzz = c8o1 * (-c2o1 + omega) * (c1o1 + c2o1 * omega) / (-c8o1 - c14o1 * omega + c7o1 * omega * omega);
- real OxyyMxzz =
- c8o1 * (-c2o1 + omega) * (-c7o1 + c4o1 * omega) / (c56o1 - c50o1 * omega + c9o1 * omega * omega);
- real Oxyz = c24o1 * (-c2o1 + omega) * (-c2o1 - c7o1 * omega + c3o1 * omega * omega) /
- (c48o1 + c152o1 * omega - c130o1 * omega * omega + c29o1 * omega * omega * omega);
- ////////////////////////////////////////////////////////////
- // 4.
- real O4 = c1o1;
- ////////////////////////////////////////////////////////////
- // 5.
- real O5 = c1o1;
- ////////////////////////////////////////////////////////////
- // 6.
- real O6 = c1o1;
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - A and DIR_00M: parameters for fourth order convergence of the diffusion term according to Eq. (115) and (116)
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a> with simplifications assuming \f$ \omega_2 = 1.0 \f$ (modify for
- //! different bulk viscosity).
- //!
- real factorA = (c4o1 + c2o1 * omega - c3o1 * omega * omega) / (c2o1 - c7o1 * omega + c5o1 * omega * omega);
- real factorB = (c4o1 + c28o1 * omega - c14o1 * omega * omega) / (c6o1 - c21o1 * omega + c15o1 * omega * omega);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Compute cumulants from central moments according to Eq. (20)-(23) in
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- ////////////////////////////////////////////////////////////
- // 4.
- real CUMcbb = mfcbb - ((mfcaa + c1o3) * mfabb + c2o1 * mfbba * mfbab) * OOrho;
- real CUMbcb = mfbcb - ((mfaca + c1o3) * mfbab + c2o1 * mfbba * mfabb) * OOrho;
- real CUMbbc = mfbbc - ((mfaac + c1o3) * mfbba + c2o1 * mfbab * mfabb) * OOrho;
-
- real CUMcca =
- mfcca - (((mfcaa * mfaca + c2o1 * mfbba * mfbba) + c1o3 * (mfcaa + mfaca)) * OOrho - c1o9 * (drho * OOrho));
- real CUMcac =
- mfcac - (((mfcaa * mfaac + c2o1 * mfbab * mfbab) + c1o3 * (mfcaa + mfaac)) * OOrho - c1o9 * (drho * OOrho));
- real CUMacc =
- mfacc - (((mfaac * mfaca + c2o1 * mfabb * mfabb) + c1o3 * (mfaac + mfaca)) * OOrho - c1o9 * (drho * OOrho));
- ////////////////////////////////////////////////////////////
- // 5.
- real CUMbcc =
- mfbcc - ((mfaac * mfbca + mfaca * mfbac + c4o1 * mfabb * mfbbb + c2o1 * (mfbab * mfacb + mfbba * mfabc)) +
- c1o3 * (mfbca + mfbac)) *
- OOrho;
- real CUMcbc =
- mfcbc - ((mfaac * mfcba + mfcaa * mfabc + c4o1 * mfbab * mfbbb + c2o1 * (mfabb * mfcab + mfbba * mfbac)) +
- c1o3 * (mfcba + mfabc)) *
- OOrho;
- real CUMccb =
- mfccb - ((mfcaa * mfacb + mfaca * mfcab + c4o1 * mfbba * mfbbb + c2o1 * (mfbab * mfbca + mfabb * mfcba)) +
- c1o3 * (mfacb + mfcab)) *
- OOrho;
- ////////////////////////////////////////////////////////////
- // 6.
- real CUMccc = mfccc + ((-c4o1 * mfbbb * mfbbb - (mfcaa * mfacc + mfaca * mfcac + mfaac * mfcca) -
- c4o1 * (mfabb * mfcbb + mfbab * mfbcb + mfbba * mfbbc) -
- c2o1 * (mfbca * mfbac + mfcba * mfabc + mfcab * mfacb)) *
- OOrho +
- (c4o1 * (mfbab * mfbab * mfaca + mfabb * mfabb * mfcaa + mfbba * mfbba * mfaac) +
- c2o1 * (mfcaa * mfaca * mfaac) + c16o1 * mfbba * mfbab * mfabb) *
- OOrho * OOrho -
- c1o3 * (mfacc + mfcac + mfcca) * OOrho - c1o9 * (mfcaa + mfaca + mfaac) * OOrho +
- (c2o1 * (mfbab * mfbab + mfabb * mfabb + mfbba * mfbba) +
- (mfaac * mfaca + mfaac * mfcaa + mfaca * mfcaa) + c1o3 * (mfaac + mfaca + mfcaa)) *
- OOrho * OOrho * c2o3 +
- c1o27 * ((drho * drho - drho) * OOrho * OOrho));
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Compute linear combinations of second and third order cumulants
- //!
- ////////////////////////////////////////////////////////////
- // 2.
- real mxxPyyPzz = mfcaa + mfaca + mfaac;
- real mxxMyy = mfcaa - mfaca;
- real mxxMzz = mfcaa - mfaac;
- ////////////////////////////////////////////////////////////
- // 3.
- real mxxyPyzz = mfcba + mfabc;
- real mxxyMyzz = mfcba - mfabc;
-
- real mxxzPyyz = mfcab + mfacb;
- real mxxzMyyz = mfcab - mfacb;
-
- real mxyyPxzz = mfbca + mfbac;
- real mxyyMxzz = mfbca - mfbac;
-
- ////////////////////////////////////////////////////////////////////////////////////
- // incl. correction
- ////////////////////////////////////////////////////////////
- //! - Compute velocity gradients from second order cumulants according to Eq. (27)-(32)
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a> Further explanations of the correction in viscosity in Appendix H of
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a> Note that the division by rho is omitted here as we need rho times
- //! the gradients later.
- //!
- real Dxy = -c3o1 * omega * mfbba;
- real Dxz = -c3o1 * omega * mfbab;
- real Dyz = -c3o1 * omega * mfabb;
- real dxux = c1o2 * (-omega) * (mxxMyy + mxxMzz) + c1o2 * OxxPyyPzz * (mfaaa - mxxPyyPzz);
- real dyuy = dxux + omega * c3o2 * mxxMyy;
- real dzuz = dxux + omega * c3o2 * mxxMzz;
- ////////////////////////////////////////////////////////////
- //! - Relaxation of second order cumulants with correction terms according to Eq. (33)-(35) in
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- mxxPyyPzz +=
- OxxPyyPzz * (mfaaa - mxxPyyPzz) - c3o1 * (c1o1 - c1o2 * OxxPyyPzz) * (vx2 * dxux + vy2 * dyuy + vz2 * dzuz);
- mxxMyy += omega * (-mxxMyy) - c3o1 * (c1o1 + c1o2 * (-omega)) * (vx2 * dxux - vy2 * dyuy);
- mxxMzz += omega * (-mxxMzz) - c3o1 * (c1o1 + c1o2 * (-omega)) * (vx2 * dxux - vz2 * dzuz);
-
- ////////////////////////////////////////////////////////////////////////////////////
- ////no correction
- // mxxPyyPzz += OxxPyyPzz*(mfaaa - mxxPyyPzz);
- // mxxMyy += -(-omega) * (-mxxMyy);
- // mxxMzz += -(-omega) * (-mxxMzz);
- //////////////////////////////////////////////////////////////////////////
- mfabb += omega * (-mfabb);
- mfbab += omega * (-mfbab);
- mfbba += omega * (-mfbba);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // relax
- //////////////////////////////////////////////////////////////////////////
- // incl. limiter
- //! - Relaxation of third order cumulants including limiter according to Eq. (116)-(123)
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- wadjust = Oxyz + (c1o1 - Oxyz) * abs(mfbbb) / (abs(mfbbb) + qudricLimitD);
- mfbbb += wadjust * (-mfbbb);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz) * abs(mxxyPyzz) / (abs(mxxyPyzz) + qudricLimitP);
- mxxyPyzz += wadjust * (-mxxyPyzz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz) * abs(mxxyMyzz) / (abs(mxxyMyzz) + qudricLimitM);
- mxxyMyzz += wadjust * (-mxxyMyzz);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz) * abs(mxxzPyyz) / (abs(mxxzPyyz) + qudricLimitP);
- mxxzPyyz += wadjust * (-mxxzPyyz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz) * abs(mxxzMyyz) / (abs(mxxzMyyz) + qudricLimitM);
- mxxzMyyz += wadjust * (-mxxzMyyz);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz) * abs(mxyyPxzz) / (abs(mxyyPxzz) + qudricLimitP);
- mxyyPxzz += wadjust * (-mxyyPxzz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz) * abs(mxyyMxzz) / (abs(mxyyMxzz) + qudricLimitM);
- mxyyMxzz += wadjust * (-mxyyMxzz);
- //////////////////////////////////////////////////////////////////////////
- // no limiter
- // mfbbb += OxyyMxzz * (-mfbbb);
- // mxxyPyzz += OxyyPxzz * (-mxxyPyzz);
- // mxxyMyzz += OxyyMxzz * (-mxxyMyzz);
- // mxxzPyyz += OxyyPxzz * (-mxxzPyyz);
- // mxxzMyyz += OxyyMxzz * (-mxxzMyyz);
- // mxyyPxzz += OxyyPxzz * (-mxyyPxzz);
- // mxyyMxzz += OxyyMxzz * (-mxyyMxzz);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Compute inverse linear combinations of second and third order cumulants
- //!
- mfcaa = c1o3 * (mxxMyy + mxxMzz + mxxPyyPzz);
- mfaca = c1o3 * (-c2o1 * mxxMyy + mxxMzz + mxxPyyPzz);
- mfaac = c1o3 * (mxxMyy - c2o1 * mxxMzz + mxxPyyPzz);
-
- mfcba = (mxxyMyzz + mxxyPyzz) * c1o2;
- mfabc = (-mxxyMyzz + mxxyPyzz) * c1o2;
- mfcab = (mxxzMyyz + mxxzPyyz) * c1o2;
- mfacb = (-mxxzMyyz + mxxzPyyz) * c1o2;
- mfbca = (mxyyMxzz + mxyyPxzz) * c1o2;
- mfbac = (-mxyyMxzz + mxyyPxzz) * c1o2;
- //////////////////////////////////////////////////////////////////////////
-
- //////////////////////////////////////////////////////////////////////////
- // 4.
- // no limiter
- //! - Relax fourth order cumulants to modified equilibrium for fourth order convergence of diffusion according
- //! to Eq. (43)-(48) <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- CUMacc = -O4 * (c1o1 / omega - c1o2) * (dyuy + dzuz) * c2o3 * factorA + (c1o1 - O4) * (CUMacc);
- CUMcac = -O4 * (c1o1 / omega - c1o2) * (dxux + dzuz) * c2o3 * factorA + (c1o1 - O4) * (CUMcac);
- CUMcca = -O4 * (c1o1 / omega - c1o2) * (dyuy + dxux) * c2o3 * factorA + (c1o1 - O4) * (CUMcca);
- CUMbbc = -O4 * (c1o1 / omega - c1o2) * Dxy * c1o3 * factorB + (c1o1 - O4) * (CUMbbc);
- CUMbcb = -O4 * (c1o1 / omega - c1o2) * Dxz * c1o3 * factorB + (c1o1 - O4) * (CUMbcb);
- CUMcbb = -O4 * (c1o1 / omega - c1o2) * Dyz * c1o3 * factorB + (c1o1 - O4) * (CUMcbb);
-
- //////////////////////////////////////////////////////////////////////////
- // 5.
- CUMbcc += O5 * (-CUMbcc);
- CUMcbc += O5 * (-CUMcbc);
- CUMccb += O5 * (-CUMccb);
-
- //////////////////////////////////////////////////////////////////////////
- // 6.
- CUMccc += O6 * (-CUMccc);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Compute central moments from post collision cumulants according to Eq. (53)-(56) in
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
-
- //////////////////////////////////////////////////////////////////////////
- // 4.
- mfcbb = CUMcbb + c1o3 * ((c3o1 * mfcaa + c1o1) * mfabb + c6o1 * mfbba * mfbab) * OOrho;
- mfbcb = CUMbcb + c1o3 * ((c3o1 * mfaca + c1o1) * mfbab + c6o1 * mfbba * mfabb) * OOrho;
- mfbbc = CUMbbc + c1o3 * ((c3o1 * mfaac + c1o1) * mfbba + c6o1 * mfbab * mfabb) * OOrho;
-
- mfcca =
- CUMcca +
- (((mfcaa * mfaca + c2o1 * mfbba * mfbba) * c9o1 + c3o1 * (mfcaa + mfaca)) * OOrho - (drho * OOrho)) * c1o9;
- mfcac =
- CUMcac +
- (((mfcaa * mfaac + c2o1 * mfbab * mfbab) * c9o1 + c3o1 * (mfcaa + mfaac)) * OOrho - (drho * OOrho)) * c1o9;
- mfacc =
- CUMacc +
- (((mfaac * mfaca + c2o1 * mfabb * mfabb) * c9o1 + c3o1 * (mfaac + mfaca)) * OOrho - (drho * OOrho)) * c1o9;
-
- //////////////////////////////////////////////////////////////////////////
- // 5.
- mfbcc = CUMbcc + c1o3 *
- (c3o1 * (mfaac * mfbca + mfaca * mfbac + c4o1 * mfabb * mfbbb +
- c2o1 * (mfbab * mfacb + mfbba * mfabc)) +
- (mfbca + mfbac)) *
- OOrho;
- mfcbc = CUMcbc + c1o3 *
- (c3o1 * (mfaac * mfcba + mfcaa * mfabc + c4o1 * mfbab * mfbbb +
- c2o1 * (mfabb * mfcab + mfbba * mfbac)) +
- (mfcba + mfabc)) *
- OOrho;
- mfccb = CUMccb + c1o3 *
- (c3o1 * (mfcaa * mfacb + mfaca * mfcab + c4o1 * mfbba * mfbbb +
- c2o1 * (mfbab * mfbca + mfabb * mfcba)) +
- (mfacb + mfcab)) *
- OOrho;
-
- //////////////////////////////////////////////////////////////////////////
- // 6.
- mfccc = CUMccc - ((-c4o1 * mfbbb * mfbbb - (mfcaa * mfacc + mfaca * mfcac + mfaac * mfcca) -
- c4o1 * (mfabb * mfcbb + mfbab * mfbcb + mfbba * mfbbc) -
- c2o1 * (mfbca * mfbac + mfcba * mfabc + mfcab * mfacb)) *
- OOrho +
- (c4o1 * (mfbab * mfbab * mfaca + mfabb * mfabb * mfcaa + mfbba * mfbba * mfaac) +
- c2o1 * (mfcaa * mfaca * mfaac) + c16o1 * mfbba * mfbab * mfabb) *
- OOrho * OOrho -
- c1o3 * (mfacc + mfcac + mfcca) * OOrho - c1o9 * (mfcaa + mfaca + mfaac) * OOrho +
- (c2o1 * (mfbab * mfbab + mfabb * mfabb + mfbba * mfbba) +
- (mfaac * mfaca + mfaac * mfcaa + mfaca * mfcaa) + c1o3 * (mfaac + mfaca + mfcaa)) *
- OOrho * OOrho * c2o3 +
- c1o27 * ((drho * drho - drho) * OOrho * OOrho));
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Add acceleration (body force) to first order cumulants according to Eq. (85)-(87) in
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a>
- //!
- mfbaa = -mfbaa;
- mfaba = -mfaba;
- mfaab = -mfaab;
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Chimera transform from central moments to well conditioned distributions as defined in Appendix J in
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a> see also Eq. (88)-(96) in <a
- //! href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017), DOI:10.1016/j.jcp.2017.05.040
- //! ]</b></a>
- //!
- ////////////////////////////////////////////////////////////////////////////////////
- // X - Dir
- backwardInverseChimeraWithK(mfaaa, mfbaa, mfcaa, vvx, vx2, c1o1, c1o1);
- backwardChimera(mfaba, mfbba, mfcba, vvx, vx2);
- backwardInverseChimeraWithK(mfaca, mfbca, mfcca, vvx, vx2, c3o1, c1o3);
- backwardChimera(mfaab, mfbab, mfcab, vvx, vx2);
- backwardChimera(mfabb, mfbbb, mfcbb, vvx, vx2);
- backwardChimera(mfacb, mfbcb, mfccb, vvx, vx2);
- backwardInverseChimeraWithK(mfaac, mfbac, mfcac, vvx, vx2, c3o1, c1o3);
- backwardChimera(mfabc, mfbbc, mfcbc, vvx, vx2);
- backwardInverseChimeraWithK(mfacc, mfbcc, mfccc, vvx, vx2, c9o1, c1o9);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // Y - Dir
- backwardInverseChimeraWithK(mfaaa, mfaba, mfaca, vvy, vy2, c6o1, c1o6);
- backwardChimera(mfaab, mfabb, mfacb, vvy, vy2);
- backwardInverseChimeraWithK(mfaac, mfabc, mfacc, vvy, vy2, c18o1, c1o18);
- backwardInverseChimeraWithK(mfbaa, mfbba, mfbca, vvy, vy2, c3o2, c2o3);
- backwardChimera(mfbab, mfbbb, mfbcb, vvy, vy2);
- backwardInverseChimeraWithK(mfbac, mfbbc, mfbcc, vvy, vy2, c9o2, c2o9);
- backwardInverseChimeraWithK(mfcaa, mfcba, mfcca, vvy, vy2, c6o1, c1o6);
- backwardChimera(mfcab, mfcbb, mfccb, vvy, vy2);
- backwardInverseChimeraWithK(mfcac, mfcbc, mfccc, vvy, vy2, c18o1, c1o18);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // Z - Dir
- backwardInverseChimeraWithK(mfaaa, mfaab, mfaac, vvz, vz2, c36o1, c1o36);
- backwardInverseChimeraWithK(mfaba, mfabb, mfabc, vvz, vz2, c9o1, c1o9);
- backwardInverseChimeraWithK(mfaca, mfacb, mfacc, vvz, vz2, c36o1, c1o36);
- backwardInverseChimeraWithK(mfbaa, mfbab, mfbac, vvz, vz2, c9o1, c1o9);
- backwardInverseChimeraWithK(mfbba, mfbbb, mfbbc, vvz, vz2, c9o4, c4o9);
- backwardInverseChimeraWithK(mfbca, mfbcb, mfbcc, vvz, vz2, c9o1, c1o9);
- backwardInverseChimeraWithK(mfcaa, mfcab, mfcac, vvz, vz2, c36o1, c1o36);
- backwardInverseChimeraWithK(mfcba, mfcbb, mfcbc, vvz, vz2, c9o1, c1o9);
- backwardInverseChimeraWithK(mfcca, mfccb, mfccc, vvz, vz2, c36o1, c1o36);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Write distributions: style of reading and writing the distributions from/to
- //! stored arrays dependent on timestep is based on the esoteric twist algorithm
- //! <a href="https://doi.org/10.3390/computation5020019"><b>[ M. Geier et al. (2017),
- //! DOI:10.3390/computation5020019 ]</b></a>
- //!
- (dist.f[DIR_P00])[k] = mfabb;
- (dist.f[DIR_M00])[kw] = mfcbb;
- (dist.f[DIR_0P0])[k] = mfbab;
- (dist.f[DIR_0M0])[ks] = mfbcb;
- (dist.f[DIR_00P])[k] = mfbba;
- (dist.f[DIR_00M])[kb] = mfbbc;
- (dist.f[DIR_PP0])[k] = mfaab;
- (dist.f[DIR_MM0])[ksw] = mfccb;
- (dist.f[DIR_PM0])[ks] = mfacb;
- (dist.f[DIR_MP0])[kw] = mfcab;
- (dist.f[DIR_P0P])[k] = mfaba;
- (dist.f[DIR_M0M])[kbw] = mfcbc;
- (dist.f[DIR_P0M])[kb] = mfabc;
- (dist.f[DIR_M0P])[kw] = mfcba;
- (dist.f[DIR_0PP])[k] = mfbaa;
- (dist.f[DIR_0MM])[kbs] = mfbcc;
- (dist.f[DIR_0PM])[kb] = mfbac;
- (dist.f[DIR_0MP])[ks] = mfbca;
- (dist.f[DIR_000])[k] = mfbbb;
- (dist.f[DIR_PPP])[k] = mfaaa;
- (dist.f[DIR_PMP])[ks] = mfaca;
- (dist.f[DIR_PPM])[kb] = mfaac;
- (dist.f[DIR_PMM])[kbs] = mfacc;
- (dist.f[DIR_MPP])[kw] = mfcaa;
- (dist.f[DIR_MMP])[ksw] = mfcca;
- (dist.f[DIR_MPM])[kbw] = mfcac;
- (dist.f[DIR_MMM])[kbsw] = mfccc;
- }
-}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream_Device.cuh b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream_Device.cuh
deleted file mode 100644
index f74192c0423ba9dc96820d7f46eecb9d49a39ed4..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream_Device.cuh
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef LB_Kernel_CUMULANT_K17_COMP_CHIM_SPARSE_H
-#define LB_Kernel_CUMULANT_K17_COMP_CHIM_SPARSE_H
-
-#include <DataTypes.h>
-#include <curand.h>
-
-__global__ void LB_Kernel_CumulantK17CompChimStream(
- real omega,
- uint* neighborX,
- uint* neighborY,
- uint* neighborZ,
- real* distributions,
- unsigned long size_Mat,
- int level,
- real* forces,
- real* quadricLimiters,
- bool isEvenTimestep,
- const uint* fluidNodeIndices,
- uint numberOfFluidNodes);
-#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim.cu
deleted file mode 100644
index a9d518d14a286ae3f6b565176969162994afa269..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim.cu
+++ /dev/null
@@ -1,57 +0,0 @@
-#include "TurbulentViscosityCumulantK17CompChim.h"
-#include "cuda/CudaGrid.h"
-#include <logger/Logger.h>
-#include "Parameter/Parameter.h"
-#include "TurbulentViscosityCumulantK17CompChim_Device.cuh"
-
-template<TurbulenceModel turbulenceModel>
-std::shared_ptr< TurbulentViscosityCumulantK17CompChim<turbulenceModel> > TurbulentViscosityCumulantK17CompChim<turbulenceModel>::getNewInstance(std::shared_ptr<Parameter> para, int level)
-{
- return std::shared_ptr<TurbulentViscosityCumulantK17CompChim<turbulenceModel> >(new TurbulentViscosityCumulantK17CompChim<turbulenceModel>(para,level));
-}
-
-template<TurbulenceModel turbulenceModel>
-void TurbulentViscosityCumulantK17CompChim<turbulenceModel>::run()
-{
- vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(para->getParH(level)->numberofthreads, para->getParH(level)->numberOfNodes);
-
- LB_Kernel_TurbulentViscosityCumulantK17CompChim < turbulenceModel > <<< grid.grid, grid.threads >>>( para->getParD(level)->omega,
- para->getParD(level)->typeOfGridNode, para->getParD(level)->neighborX,
- para->getParD(level)->neighborY,
- para->getParD(level)->neighborZ,
- para->getParD(level)->distributions.f[0],
- para->getParD(level)->rho,
- para->getParD(level)->velocityX,
- para->getParD(level)->velocityY,
- para->getParD(level)->velocityZ,
- para->getParD(level)->turbViscosity,
- para->getSGSConstant(),
- (unsigned long)para->getParD(level)->numberOfNodes,
- level,
- para->getIsBodyForce(),
- para->getForcesDev(),
- para->getParD(level)->forceX_SP,
- para->getParD(level)->forceY_SP,
- para->getParD(level)->forceZ_SP,
- para->getQuadricLimitersDev(),
- para->getParD(level)->isEvenTimestep);
-
- getLastCudaError("LB_Kernel_TurbulentViscosityCumulantK17CompChim execution failed");
-}
-
-template<TurbulenceModel turbulenceModel>
-TurbulentViscosityCumulantK17CompChim<turbulenceModel>::TurbulentViscosityCumulantK17CompChim(std::shared_ptr<Parameter> para, int level)
-{
- this->para = para;
- this->level = level;
-
- myPreProcessorTypes.push_back(InitCompSP27);
-
- myKernelGroup = BasicKernel;
-
- VF_LOG_INFO("Using turbulence model: {}", turbulenceModel);
-}
-
-template class TurbulentViscosityCumulantK17CompChim<TurbulenceModel::AMD>;
-template class TurbulentViscosityCumulantK17CompChim<TurbulenceModel::Smagorinsky>;
-template class TurbulentViscosityCumulantK17CompChim<TurbulenceModel::QR>;
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim.h b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim.h
deleted file mode 100644
index 0d35b68c916e54c6ec6eeeacd7189fe4d9a33c10..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim.h
+++ /dev/null
@@ -1,19 +0,0 @@
-#ifndef TurbulentViscosityCUMULANT_K17_COMP_CHIM_H
-#define TurbulentViscosityCUMULANT_K17_COMP_CHIM_H
-
-#include "Kernel/KernelImp.h"
-#include "Parameter/Parameter.h"
-
-template<TurbulenceModel turbulenceModel>
-class TurbulentViscosityCumulantK17CompChim : public KernelImp
-{
-public:
- static std::shared_ptr< TurbulentViscosityCumulantK17CompChim<turbulenceModel> > getNewInstance(std::shared_ptr< Parameter> para, int level);
- void run();
-
-private:
- TurbulentViscosityCumulantK17CompChim();
- TurbulentViscosityCumulantK17CompChim(std::shared_ptr<Parameter> para, int level);
-};
-
-#endif
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim_Device.cu b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim_Device.cu
deleted file mode 100644
index 32350b95107b68103af0f238fefe095882919092..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim_Device.cu
+++ /dev/null
@@ -1,687 +0,0 @@
-//=======================================================================================
-// ____ ____ __ ______ __________ __ __ __ __
-// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
-// \ \ | | | | | |_) | | | | | | | / \ | |
-// \ \ | | | | | _ / | | | | | | / /\ \ | |
-// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
-// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
-// \ \ | | ________________________________________________________________
-// \ \ | | | ______________________________________________________________|
-// \ \| | | | __ __ __ __ ______ _______
-// \ | | |_____ | | | | | | | | | _ \ / _____)
-// \ | | _____| | | | | | | | | | | \ \ \_______
-// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
-// \ _____| |__| |________| \_______/ |__| |______/ (_______/
-//
-// This file is part of VirtualFluids. VirtualFluids is free software: you can
-// redistribute it and/or modify it under the terms of the GNU General Public
-// License as published by the Free Software Foundation, either version 3 of
-// the License, or (at your option) any later version.
-//
-// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
-// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
-// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
-// for more details.
-//
-// You should have received a copy of the GNU General Public License along
-// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
-//
-//! \file TurbulentViscosityCumulantK17CompChim_Device.cu
-//! \author Henry Korb, Henrik Asmuth
-//! \date 16/05/2022
-//! \brief CumulantK17CompChim kernel by Martin Schönherr that inlcudes turbulent viscosity and other small mods.
-//!
-//! Additions to CumulantK17CompChim:
-//! - can incorporate local body force
-//! - when applying a local body force, the total round of error of forcing+bodyforce is saved and added in next time step
-//! - uses turbulent viscosity that is computed in separate kernel (as of now AMD)
-//! - saves macroscopic values (needed for instance for probes, AMD, and actuator models)
-//!
-//=======================================================================================
-/* Device code */
-#include "LBM/LB.h"
-#include "lbm/constants/D3Q27.h"
-#include <lbm/constants/NumericConstants.h>
-#include "Kernel/Utilities/DistributionHelper.cuh"
-
-#include "GPU/TurbulentViscosityInlines.cuh"
-
-using namespace vf::lbm::constant;
-using namespace vf::lbm::dir;
-#include "Kernel/Utilities/ChimeraTransformation.h"
-
-
-////////////////////////////////////////////////////////////////////////////////
-template<TurbulenceModel turbulenceModel>
-__global__ void LB_Kernel_TurbulentViscosityCumulantK17CompChim(
- real omega_in,
- uint* typeOfGridNode,
- uint* neighborX,
- uint* neighborY,
- uint* neighborZ,
- real* distributions,
- real* rho,
- real* vx,
- real* vy,
- real* vz,
- real* turbulentViscosity,
- real SGSconstant,
- unsigned long size_Mat,
- int level,
- bool bodyForce,
- real* forces,
- real* bodyForceX,
- real* bodyForceY,
- real* bodyForceZ,
- real* quadricLimiters,
- bool isEvenTimestep)
-{
- //////////////////////////////////////////////////////////////////////////
- //! Cumulant K17 Kernel is based on \ref
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017), DOI:10.1016/j.jcp.2017.05.040
- //! ]</b></a> and \ref <a href="https://doi.org/10.1016/j.jcp.2017.07.004"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.07.004 ]</b></a>
- //!
- //! The cumulant kernel is executed in the following steps
- //!
- ////////////////////////////////////////////////////////////////////////////////
- //! - Get node index coordinates from threadIdx, blockIdx, blockDim and gridDim.
- //!
- const unsigned k_000 = vf::gpu::getNodeIndex();
-
- //////////////////////////////////////////////////////////////////////////
- // run for all indices in size_Mat and fluid nodes
- if ((k_000 < size_Mat) && (typeOfGridNode[k_000] == GEO_FLUID)) {
- //////////////////////////////////////////////////////////////////////////
- //! - Read distributions: style of reading and writing the distributions from/to stored arrays dependent on
- //! timestep is based on the esoteric twist algorithm \ref <a
- //! href="https://doi.org/10.3390/computation5020019"><b>[ M. Geier et al. (2017),
- //! DOI:10.3390/computation5020019 ]</b></a>
- //!
- Distributions27 dist = vf::gpu::getDistributionReferences27(distributions, size_Mat, isEvenTimestep);
-
- ////////////////////////////////////////////////////////////////////////////////
- //! - Set neighbor indices (necessary for indirect addressing)
- uint k_M00 = neighborX[k_000];
- uint k_0M0 = neighborY[k_000];
- uint k_00M = neighborZ[k_000];
- uint k_MM0 = neighborY[k_M00];
- uint k_M0M = neighborZ[k_M00];
- uint k_0MM = neighborZ[k_0M0];
- uint k_MMM = neighborZ[k_MM0];
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Set local distributions
- //!
- real f_000 = (dist.f[DIR_000])[k_000];
- real f_P00 = (dist.f[DIR_P00])[k_000];
- real f_M00 = (dist.f[DIR_M00])[k_M00];
- real f_0P0 = (dist.f[DIR_0P0])[k_000];
- real f_0M0 = (dist.f[DIR_0M0])[k_0M0];
- real f_00P = (dist.f[DIR_00P])[k_000];
- real f_00M = (dist.f[DIR_00M])[k_00M];
- real f_PP0 = (dist.f[DIR_PP0])[k_000];
- real f_MM0 = (dist.f[DIR_MM0])[k_MM0];
- real f_PM0 = (dist.f[DIR_PM0])[k_0M0];
- real f_MP0 = (dist.f[DIR_MP0])[k_M00];
- real f_P0P = (dist.f[DIR_P0P])[k_000];
- real f_M0M = (dist.f[DIR_M0M])[k_M0M];
- real f_P0M = (dist.f[DIR_P0M])[k_00M];
- real f_M0P = (dist.f[DIR_M0P])[k_M00];
- real f_0PP = (dist.f[DIR_0PP])[k_000];
- real f_0MM = (dist.f[DIR_0MM])[k_0MM];
- real f_0PM = (dist.f[DIR_0PM])[k_00M];
- real f_0MP = (dist.f[DIR_0MP])[k_0M0];
- real f_PPP = (dist.f[DIR_PPP])[k_000];
- real f_MPP = (dist.f[DIR_MPP])[k_M00];
- real f_PMP = (dist.f[DIR_PMP])[k_0M0];
- real f_MMP = (dist.f[DIR_MMP])[k_MM0];
- real f_PPM = (dist.f[DIR_PPM])[k_00M];
- real f_MPM = (dist.f[DIR_MPM])[k_M0M];
- real f_PMM = (dist.f[DIR_PMM])[k_0MM];
- real f_MMM = (dist.f[DIR_MMM])[k_MMM];
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Define aliases to use the same variable for the moments (m's):
- //!
- real& m_111 = f_000;
- real& m_211 = f_P00;
- real& m_011 = f_M00;
- real& m_121 = f_0P0;
- real& m_101 = f_0M0;
- real& m_112 = f_00P;
- real& m_110 = f_00M;
- real& m_221 = f_PP0;
- real& m_001 = f_MM0;
- real& m_201 = f_PM0;
- real& m_021 = f_MP0;
- real& m_212 = f_P0P;
- real& m_010 = f_M0M;
- real& m_210 = f_P0M;
- real& m_012 = f_M0P;
- real& m_122 = f_0PP;
- real& m_100 = f_0MM;
- real& m_120 = f_0PM;
- real& m_102 = f_0MP;
- real& m_222 = f_PPP;
- real& m_022 = f_MPP;
- real& m_202 = f_PMP;
- real& m_002 = f_MMP;
- real& m_220 = f_PPM;
- real& m_020 = f_MPM;
- real& m_200 = f_PMM;
- real& m_000 = f_MMM;
-
- //////////////////////////////////////////////////////(unsigned long)//////////////////////////////
- //! - Calculate density and velocity using pyramid summation for low round-off errors as in Eq. (J1)-(J3) \ref
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a>
- //!
- real drho = ((((f_PPP + f_MMM) + (f_MPM + f_PMP)) + ((f_MPP + f_PMM) + (f_MMP + f_PPM))) +
- (((f_0MP + f_0PM) + (f_0MM + f_0PP)) + ((f_M0P + f_P0M) + (f_M0M + f_P0P)) +
- ((f_MP0 + f_PM0) + (f_MM0 + f_PP0))) +
- ((f_M00 + f_P00) + (f_0M0 + f_0P0) + (f_00M + f_00P))) +
- f_000;
-
- real oneOverRho = c1o1 / (c1o1 + drho);
-
- real vvx = ((((f_PPP - f_MMM) + (f_PMP - f_MPM)) + ((f_PMM - f_MPP) + (f_PPM - f_MMP))) +
- (((f_P0M - f_M0P) + (f_P0P - f_M0M)) + ((f_PM0 - f_MP0) + (f_PP0 - f_MM0))) + (f_P00 - f_M00)) *
- oneOverRho;
- real vvy = ((((f_PPP - f_MMM) + (f_MPM - f_PMP)) + ((f_MPP - f_PMM) + (f_PPM - f_MMP))) +
- (((f_0PM - f_0MP) + (f_0PP - f_0MM)) + ((f_MP0 - f_PM0) + (f_PP0 - f_MM0))) + (f_0P0 - f_0M0)) *
- oneOverRho;
- real vvz = ((((f_PPP - f_MMM) + (f_PMP - f_MPM)) + ((f_MPP - f_PMM) + (f_MMP - f_PPM))) +
- (((f_0MP - f_0PM) + (f_0PP - f_0MM)) + ((f_M0P - f_P0M) + (f_P0P - f_M0M))) + (f_00P - f_00M)) *
- oneOverRho;
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Add half of the acceleration (body force) to the velocity as in Eq. (42) \ref
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a>
- //!
- real factor = c1o1;
- for (size_t i = 1; i <= level; i++) {
- factor *= c2o1;
- }
-
- real fx = forces[0];
- real fy = forces[1];
- real fz = forces[2];
-
- if( bodyForce ){
- fx += bodyForceX[k_000];
- fy += bodyForceY[k_000];
- fz += bodyForceZ[k_000];
-
- real vx = vvx;
- real vy = vvy;
- real vz = vvz;
- real acc_x = fx * c1o2 / factor;
- real acc_y = fy * c1o2 / factor;
- real acc_z = fz * c1o2 / factor;
-
- vvx += acc_x;
- vvy += acc_y;
- vvz += acc_z;
-
- // // Reset body force. To be used when not using round-off correction.
- // bodyForceX[k] = 0.0f;
- // bodyForceY[k] = 0.0f;
- // bodyForceZ[k] = 0.0f;
-
- ////////////////////////////////////////////////////////////////////////////////////
- //!> Round-off correction
- //!
- //!> Similar to Kahan summation algorithm (https://en.wikipedia.org/wiki/Kahan_summation_algorithm)
- //!> Essentially computes the round-off error of the applied force and adds it in the next time step as a compensation.
- //!> Seems to be necesseary at very high Re boundary layers, where the forcing and velocity can
- //!> differ by several orders of magnitude.
- //!> \note 16/05/2022: Testing, still ongoing!
- //!
- bodyForceX[k_000] = (acc_x-(vvx-vx))*factor*c2o1;
- bodyForceY[k_000] = (acc_y-(vvy-vy))*factor*c2o1;
- bodyForceZ[k_000] = (acc_z-(vvz-vz))*factor*c2o1;
- }
- else{
- vvx += fx * c1o2 / factor;
- vvy += fy * c1o2 / factor;
- vvz += fz * c1o2 / factor;
- }
-
-
- ////////////////////////////////////////////////////////////////////////////////////
- // calculate the square of velocities for this lattice node
- real vx2 = vvx * vvx;
- real vy2 = vvy * vvy;
- real vz2 = vvz * vvz;
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Set relaxation limiters for third order cumulants to default value \f$ \lambda=0.001 \f$ according to
- //! section 6 in \ref <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- real quadricLimitP = quadricLimiters[0];
- real quadricLimitM = quadricLimiters[1];
- real quadricLimitD = quadricLimiters[2];
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Chimera transform from well conditioned distributions to central moments as defined in Appendix J in \ref
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a> see also Eq. (6)-(14) in \ref <a
- //! href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017), DOI:10.1016/j.jcp.2017.05.040
- //! ]</b></a>
- //!
- ////////////////////////////////////////////////////////////////////////////////////
- // Z - Dir
- forwardInverseChimeraWithK(f_MMM, f_MM0, f_MMP, vvz, vz2, c36o1, c1o36);
- forwardInverseChimeraWithK(f_M0M, f_M00, f_M0P, vvz, vz2, c9o1, c1o9);
- forwardInverseChimeraWithK(f_MPM, f_MP0, f_MPP, vvz, vz2, c36o1, c1o36);
- forwardInverseChimeraWithK(f_0MM, f_0M0, f_0MP, vvz, vz2, c9o1, c1o9);
- forwardInverseChimeraWithK(f_00M, f_000, f_00P, vvz, vz2, c9o4, c4o9);
- forwardInverseChimeraWithK(f_0PM, f_0P0, f_0PP, vvz, vz2, c9o1, c1o9);
- forwardInverseChimeraWithK(f_PMM, f_PM0, f_PMP, vvz, vz2, c36o1, c1o36);
- forwardInverseChimeraWithK(f_P0M, f_P00, f_P0P, vvz, vz2, c9o1, c1o9);
- forwardInverseChimeraWithK(f_PPM, f_PP0, f_PPP, vvz, vz2, c36o1, c1o36);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // Y - Dir
- forwardInverseChimeraWithK(f_MMM, f_M0M, f_MPM, vvy, vy2, c6o1, c1o6);
- forwardChimera( f_MM0, f_M00, f_MP0, vvy, vy2);
- forwardInverseChimeraWithK(f_MMP, f_M0P, f_MPP, vvy, vy2, c18o1, c1o18);
- forwardInverseChimeraWithK(f_0MM, f_00M, f_0PM, vvy, vy2, c3o2, c2o3);
- forwardChimera( f_0M0, f_000, f_0P0, vvy, vy2);
- forwardInverseChimeraWithK(f_0MP, f_00P, f_0PP, vvy, vy2, c9o2, c2o9);
- forwardInverseChimeraWithK(f_PMM, f_P0M, f_PPM, vvy, vy2, c6o1, c1o6);
- forwardChimera( f_PM0, f_P00, f_PP0, vvy, vy2);
- forwardInverseChimeraWithK(f_PMP, f_P0P, f_PPP, vvy, vy2, c18o1, c1o18);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // X - Dir
- forwardInverseChimeraWithK(f_MMM, f_0MM, f_PMM, vvx, vx2, c1o1, c1o1);
- forwardChimera( f_M0M, f_00M, f_P0M, vvx, vx2);
- forwardInverseChimeraWithK(f_MPM, f_0PM, f_PPM, vvx, vx2, c3o1, c1o3);
- forwardChimera( f_MM0, f_0M0, f_PM0, vvx, vx2);
- forwardChimera( f_M00, f_000, f_P00, vvx, vx2);
- forwardChimera( f_MP0, f_0P0, f_PP0, vvx, vx2);
- forwardInverseChimeraWithK(f_MMP, f_0MP, f_PMP, vvx, vx2, c3o1, c1o3);
- forwardChimera( f_M0P, f_00P, f_P0P, vvx, vx2);
- forwardInverseChimeraWithK(f_MPP, f_0PP, f_PPP, vvx, vx2, c3o1, c1o9);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Setting relaxation rates for non-hydrodynamic cumulants (default values). Variable names and equations
- //! according to <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //! => [NAME IN PAPER]=[NAME IN CODE]=[DEFAULT VALUE].
- //! - Trace of second order cumulants \f$ C_{200}+C_{020}+C_{002} \f$ used to adjust bulk
- //! viscosity:\f$\omega_2=OxxPyyPzz=1.0 \f$.
- //! - Third order cumulants \f$ C_{120}+C_{102}, C_{210}+C_{012}, C_{201}+C_{021} \f$: \f$ \omega_3=OxyyPxzz
- //! \f$ set according to Eq. (111) with simplifications assuming \f$ \omega_2=1.0\f$.
- //! - Third order cumulants \f$ C_{120}-C_{102}, C_{210}-C_{012}, C_{201}-C_{021} \f$: \f$ \omega_4 = OxyyMxzz
- //! \f$ set according to Eq. (112) with simplifications assuming \f$ \omega_2 = 1.0\f$.
- //! - Third order cumulants \f$ C_{111} \f$: \f$ \omega_5 = Oxyz \f$ set according to Eq. (113) with
- //! simplifications assuming \f$ \omega_2 = 1.0\f$ (modify for different bulk viscosity).
- //! - Fourth order cumulants \f$ C_{220}, C_{202}, C_{022}, C_{211}, C_{121}, C_{112} \f$: for simplification
- //! all set to the same default value \f$ \omega_6=\omega_7=\omega_8=O4=1.0 \f$.
- //! - Fifth order cumulants \f$ C_{221}, C_{212}, C_{122}\f$: \f$\omega_9=O5=1.0\f$.
- //! - Sixth order cumulant \f$ C_{222}\f$: \f$\omega_{10}=O6=1.0\f$.
- //!
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Calculate modified omega with turbulent viscosity
- //!
- real omega = omega_in / (c1o1 + c3o1*omega_in*turbulentViscosity[k_000]);
- ////////////////////////////////////////////////////////////
- // 2.
- real OxxPyyPzz = c1o1;
- ////////////////////////////////////////////////////////////
- // 3.
- real OxyyPxzz = c8o1 * (-c2o1 + omega) * (c1o1 + c2o1 * omega) / (-c8o1 - c14o1 * omega + c7o1 * omega * omega);
- real OxyyMxzz =
- c8o1 * (-c2o1 + omega) * (-c7o1 + c4o1 * omega) / (c56o1 - c50o1 * omega + c9o1 * omega * omega);
- real Oxyz = c24o1 * (-c2o1 + omega) * (-c2o1 - c7o1 * omega + c3o1 * omega * omega) /
- (c48o1 + c152o1 * omega - c130o1 * omega * omega + c29o1 * omega * omega * omega);
- ////////////////////////////////////////////////////////////
- // 4.
- real O4 = c1o1;
- ////////////////////////////////////////////////////////////
- // 5.
- real O5 = c1o1;
- ////////////////////////////////////////////////////////////
- // 6.
- real O6 = c1o1;
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - A and DIR_00M: parameters for fourth order convergence of the diffusion term according to Eq. (115) and (116)
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a> with simplifications assuming \f$ \omega_2 = 1.0 \f$ (modify for
- //! different bulk viscosity).
- //!
- real factorA = (c4o1 + c2o1 * omega - c3o1 * omega * omega) / (c2o1 - c7o1 * omega + c5o1 * omega * omega);
- real factorB = (c4o1 + c28o1 * omega - c14o1 * omega * omega) / (c6o1 - c21o1 * omega + c15o1 * omega * omega);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Compute cumulants from central moments according to Eq. (20)-(23) in
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- ////////////////////////////////////////////////////////////
- // 4.
- real c_211 = m_211 - ((m_200 + c1o3) * m_011 + c2o1 * m_110 * m_101) * oneOverRho;
- real c_121 = m_121 - ((m_020 + c1o3) * m_101 + c2o1 * m_110 * m_011) * oneOverRho;
- real c_112 = m_112 - ((m_002 + c1o3) * m_110 + c2o1 * m_101 * m_011) * oneOverRho;
-
- real c_220 = m_220 - (((m_200 * m_020 + c2o1 * m_110 * m_110) + c1o3 * (m_200 + m_020)) * oneOverRho - c1o9 * (drho * oneOverRho));
- real c_202 = m_202 - (((m_200 * m_002 + c2o1 * m_101 * m_101) + c1o3 * (m_200 + m_002)) * oneOverRho - c1o9 * (drho * oneOverRho));
- real c_022 = m_022 - (((m_002 * m_020 + c2o1 * m_011 * m_011) + c1o3 * (m_002 + m_020)) * oneOverRho - c1o9 * (drho * oneOverRho));
- ////////////////////////////////////////////////////////////
- // 5.
- real c_122 =
- m_122 - ((m_002 * m_120 + m_020 * m_102 + c4o1 * m_011 * m_111 + c2o1 * (m_101 * m_021 + m_110 * m_012)) +
- c1o3 * (m_120 + m_102)) *
- oneOverRho;
- real c_212 =
- m_212 - ((m_002 * m_210 + m_200 * m_012 + c4o1 * m_101 * m_111 + c2o1 * (m_011 * m_201 + m_110 * m_102)) +
- c1o3 * (m_210 + m_012)) *
- oneOverRho;
- real c_221 =
- m_221 - ((m_200 * m_021 + m_020 * m_201 + c4o1 * m_110 * m_111 + c2o1 * (m_101 * m_120 + m_011 * m_210)) +
- c1o3 * (m_021 + m_201)) *
- oneOverRho;
- ////////////////////////////////////////////////////////////
- // 6.
- real c_222 = m_222 + ((-c4o1 * m_111 * m_111 - (m_200 * m_022 + m_020 * m_202 + m_002 * m_220) -
- c4o1 * (m_011 * m_211 + m_101 * m_121 + m_110 * m_112) -
- c2o1 * (m_120 * m_102 + m_210 * m_012 + m_201 * m_021)) *
- oneOverRho +
- (c4o1 * (m_101 * m_101 * m_020 + m_011 * m_011 * m_200 + m_110 * m_110 * m_002) +
- c2o1 * (m_200 * m_020 * m_002) + c16o1 * m_110 * m_101 * m_011) *
- oneOverRho * oneOverRho -
- c1o3 * (m_022 + m_202 + m_220) * oneOverRho - c1o9 * (m_200 + m_020 + m_002) * oneOverRho +
- (c2o1 * (m_101 * m_101 + m_011 * m_011 + m_110 * m_110) +
- (m_002 * m_020 + m_002 * m_200 + m_020 * m_200) + c1o3 * (m_002 + m_020 + m_200)) *
- oneOverRho * oneOverRho * c2o3 +
- c1o27 * ((drho * drho - drho) * oneOverRho * oneOverRho));
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Compute linear combinations of second and third order cumulants
- //!
- ////////////////////////////////////////////////////////////
- // 2.
- real mxxPyyPzz = m_200 + m_020 + m_002;
- real mxxMyy = m_200 - m_020;
- real mxxMzz = m_200 - m_002;
- ////////////////////////////////////////////////////////////
- // 3.
- real mxxyPyzz = m_210 + m_012;
- real mxxyMyzz = m_210 - m_012;
-
- real mxxzPyyz = m_201 + m_021;
- real mxxzMyyz = m_201 - m_021;
-
- real mxyyPxzz = m_120 + m_102;
- real mxyyMxzz = m_120 - m_102;
-
- ////////////////////////////////////////////////////////////////////////////////////
- // incl. correction
- ////////////////////////////////////////////////////////////
- //! - Compute velocity gradients from second order cumulants according to Eq. (27)-(32)
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a> Further explanations of the correction in viscosity in Appendix H of
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a> Note that the division by rho is omitted here as we need rho times
- //! the gradients later.
- //!
- real Dxy = -c3o1 * omega * m_110;
- real Dxz = -c3o1 * omega * m_101;
- real Dyz = -c3o1 * omega * m_011;
- real dxux = c1o2 * (-omega) * (mxxMyy + mxxMzz) + c1o2 * OxxPyyPzz * (m_000 - mxxPyyPzz);
- real dyuy = dxux + omega * c3o2 * mxxMyy;
- real dzuz = dxux + omega * c3o2 * mxxMzz;
-
- ////////////////////////////////////////////////////////////////////////////////////
- switch (turbulenceModel)
- {
- case TurbulenceModel::AMD: //AMD is computed in separate kernel
- break;
- case TurbulenceModel::Smagorinsky:
- turbulentViscosity[k_000] = calcTurbulentViscositySmagorinsky(SGSconstant, dxux, dyuy, dzuz, Dxy, Dxz , Dyz);
- break;
- case TurbulenceModel::QR:
- turbulentViscosity[k_000] = calcTurbulentViscosityQR(SGSconstant, dxux, dyuy, dzuz, Dxy, Dxz , Dyz);
- break;
- default:
- break;
- }
- ////////////////////////////////////////////////////////////
- //! - Relaxation of second order cumulants with correction terms according to Eq. (33)-(35) in
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- mxxPyyPzz += OxxPyyPzz * (m_000 - mxxPyyPzz) - c3o1 * (c1o1 - c1o2 * OxxPyyPzz) * (vx2 * dxux + vy2 * dyuy + vz2 * dzuz);
- mxxMyy += omega * (-mxxMyy) - c3o1 * (c1o1 + c1o2 * (-omega)) * (vx2 * dxux - vy2 * dyuy);
- mxxMzz += omega * (-mxxMzz) - c3o1 * (c1o1 + c1o2 * (-omega)) * (vx2 * dxux - vz2 * dzuz);
-
- ////////////////////////////////////////////////////////////////////////////////////
- ////no correction
- // mxxPyyPzz += OxxPyyPzz*(mfaaa - mxxPyyPzz);
- // mxxMyy += -(-omega) * (-mxxMyy);
- // mxxMzz += -(-omega) * (-mxxMzz);
- //////////////////////////////////////////////////////////////////////////
- m_011 += omega * (-m_011);
- m_101 += omega * (-m_101);
- m_110 += omega * (-m_110);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // relax
- //////////////////////////////////////////////////////////////////////////
- // incl. limiter
- //! - Relaxation of third order cumulants including limiter according to Eq. (116)-(123)
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- real wadjust = Oxyz + (c1o1 - Oxyz) * abs(m_111) / (abs(m_111) + quadricLimitD);
- m_111 += wadjust * (-m_111);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz) * abs(mxxyPyzz) / (abs(mxxyPyzz) + quadricLimitP);
- mxxyPyzz += wadjust * (-mxxyPyzz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz) * abs(mxxyMyzz) / (abs(mxxyMyzz) + quadricLimitM);
- mxxyMyzz += wadjust * (-mxxyMyzz);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz) * abs(mxxzPyyz) / (abs(mxxzPyyz) + quadricLimitP);
- mxxzPyyz += wadjust * (-mxxzPyyz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz) * abs(mxxzMyyz) / (abs(mxxzMyyz) + quadricLimitM);
- mxxzMyyz += wadjust * (-mxxzMyyz);
- wadjust = OxyyPxzz + (c1o1 - OxyyPxzz) * abs(mxyyPxzz) / (abs(mxyyPxzz) + quadricLimitP);
- mxyyPxzz += wadjust * (-mxyyPxzz);
- wadjust = OxyyMxzz + (c1o1 - OxyyMxzz) * abs(mxyyMxzz) / (abs(mxyyMxzz) + quadricLimitM);
- mxyyMxzz += wadjust * (-mxyyMxzz);
- //////////////////////////////////////////////////////////////////////////
- // no limiter
- // mfbbb += OxyyMxzz * (-mfbbb);
- // mxxyPyzz += OxyyPxzz * (-mxxyPyzz);
- // mxxyMyzz += OxyyMxzz * (-mxxyMyzz);
- // mxxzPyyz += OxyyPxzz * (-mxxzPyyz);
- // mxxzMyyz += OxyyMxzz * (-mxxzMyyz);
- // mxyyPxzz += OxyyPxzz * (-mxyyPxzz);
- // mxyyMxzz += OxyyMxzz * (-mxyyMxzz);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Compute inverse linear combinations of second and third order cumulants
- //!
- m_200 = c1o3 * (mxxMyy + mxxMzz + mxxPyyPzz);
- m_020 = c1o3 * (-c2o1 * mxxMyy + mxxMzz + mxxPyyPzz);
- m_002 = c1o3 * (mxxMyy - c2o1 * mxxMzz + mxxPyyPzz);
-
- m_210 = ( mxxyMyzz + mxxyPyzz) * c1o2;
- m_012 = (-mxxyMyzz + mxxyPyzz) * c1o2;
- m_201 = ( mxxzMyyz + mxxzPyyz) * c1o2;
- m_021 = (-mxxzMyyz + mxxzPyyz) * c1o2;
- m_120 = ( mxyyMxzz + mxyyPxzz) * c1o2;
- m_102 = (-mxyyMxzz + mxyyPxzz) * c1o2;
- //////////////////////////////////////////////////////////////////////////
-
- //////////////////////////////////////////////////////////////////////////
- // 4.
- // no limiter
- //! - Relax fourth order cumulants to modified equilibrium for fourth order convergence of diffusion according
- //! to Eq. (43)-(48) <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
- c_022 = -O4 * (c1o1 / omega - c1o2) * (dyuy + dzuz) * c2o3 * factorA + (c1o1 - O4) * (c_022);
- c_202 = -O4 * (c1o1 / omega - c1o2) * (dxux + dzuz) * c2o3 * factorA + (c1o1 - O4) * (c_202);
- c_220 = -O4 * (c1o1 / omega - c1o2) * (dyuy + dxux) * c2o3 * factorA + (c1o1 - O4) * (c_220);
- c_112 = -O4 * (c1o1 / omega - c1o2) * Dxy * c1o3 * factorB + (c1o1 - O4) * (c_112);
- c_121 = -O4 * (c1o1 / omega - c1o2) * Dxz * c1o3 * factorB + (c1o1 - O4) * (c_121);
- c_211 = -O4 * (c1o1 / omega - c1o2) * Dyz * c1o3 * factorB + (c1o1 - O4) * (c_211);
-
-
- //////////////////////////////////////////////////////////////////////////
- // 5.
- c_122 += O5 * (-c_122);
- c_212 += O5 * (-c_212);
- c_221 += O5 * (-c_221);
-
- //////////////////////////////////////////////////////////////////////////
- // 6.
- c_222 += O6 * (-c_222);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Compute central moments from post collision cumulants according to Eq. (53)-(56) in
- //! <a href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017),
- //! DOI:10.1016/j.jcp.2017.05.040 ]</b></a>
- //!
-
- //////////////////////////////////////////////////////////////////////////
- // 4.
- m_211 = c_211 + c1o3 * ((c3o1 * m_200 + c1o1) * m_011 + c6o1 * m_110 * m_101) * oneOverRho;
- m_121 = c_121 + c1o3 * ((c3o1 * m_020 + c1o1) * m_101 + c6o1 * m_110 * m_011) * oneOverRho;
- m_112 = c_112 + c1o3 * ((c3o1 * m_002 + c1o1) * m_110 + c6o1 * m_101 * m_011) * oneOverRho;
-
- m_220 =
- c_220 + (((m_200 * m_020 + c2o1 * m_110 * m_110) * c9o1 + c3o1 * (m_200 + m_020)) * oneOverRho - (drho * oneOverRho)) * c1o9;
- m_202 =
- c_202 + (((m_200 * m_002 + c2o1 * m_101 * m_101) * c9o1 + c3o1 * (m_200 + m_002)) * oneOverRho - (drho * oneOverRho)) * c1o9;
- m_022 =
- c_022 + (((m_002 * m_020 + c2o1 * m_011 * m_011) * c9o1 + c3o1 * (m_002 + m_020)) * oneOverRho - (drho * oneOverRho)) * c1o9;
-
- //////////////////////////////////////////////////////////////////////////
- // 5.
- m_122 = c_122 + c1o3 *
- (c3o1 * (m_002 * m_120 + m_020 * m_102 + c4o1 * m_011 * m_111 + c2o1 * (m_101 * m_021 + m_110 * m_012)) +
- (m_120 + m_102)) * oneOverRho;
- m_212 = c_212 + c1o3 *
- (c3o1 * (m_002 * m_210 + m_200 * m_012 + c4o1 * m_101 * m_111 + c2o1 * (m_011 * m_201 + m_110 * m_102)) +
- (m_210 + m_012)) * oneOverRho;
- m_221 = c_221 + c1o3 *
- (c3o1 * (m_200 * m_021 + m_020 * m_201 + c4o1 * m_110 * m_111 + c2o1 * (m_101 * m_120 + m_011 * m_210)) +
- (m_021 + m_201)) * oneOverRho;
-
- //////////////////////////////////////////////////////////////////////////
- // 6.
- m_222 = c_222 - ((-c4o1 * m_111 * m_111 - (m_200 * m_022 + m_020 * m_202 + m_002 * m_220) -
- c4o1 * (m_011 * m_211 + m_101 * m_121 + m_110 * m_112) -
- c2o1 * (m_120 * m_102 + m_210 * m_012 + m_201 * m_021)) *
- oneOverRho +
- (c4o1 * (m_101 * m_101 * m_020 + m_011 * m_011 * m_200 + m_110 * m_110 * m_002) +
- c2o1 * (m_200 * m_020 * m_002) + c16o1 * m_110 * m_101 * m_011) *
- oneOverRho * oneOverRho -
- c1o3 * (m_022 + m_202 + m_220) * oneOverRho - c1o9 * (m_200 + m_020 + m_002) * oneOverRho +
- (c2o1 * (m_101 * m_101 + m_011 * m_011 + m_110 * m_110) +
- (m_002 * m_020 + m_002 * m_200 + m_020 * m_200) + c1o3 * (m_002 + m_020 + m_200)) *
- oneOverRho * oneOverRho * c2o3 +
- c1o27 * ((drho * drho - drho) * oneOverRho * oneOverRho));
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Add acceleration (body force) to first order cumulants according to Eq. (85)-(87) in
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a>
- //!
- m_100 = -m_100;
- m_010 = -m_010;
- m_001 = -m_001;
-
- //Write to array here to distribute read/write
- rho[k_000] = drho;
- vx[k_000] = vvx;
- vy[k_000] = vvy;
- vz[k_000] = vvz;
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Chimera transform from central moments to well conditioned distributions as defined in Appendix J in
- //! <a href="https://doi.org/10.1016/j.camwa.2015.05.001"><b>[ M. Geier et al. (2015),
- //! DOI:10.1016/j.camwa.2015.05.001 ]</b></a> see also Eq. (88)-(96) in <a
- //! href="https://doi.org/10.1016/j.jcp.2017.05.040"><b>[ M. Geier et al. (2017), DOI:10.1016/j.jcp.2017.05.040
- //! ]</b></a>
- //!
- ////////////////////////////////////////////////////////////////////////////////////
- // X - Dir
- backwardInverseChimeraWithK(m_000, m_100, m_200, vvx, vx2, c1o1, c1o1);
- backwardChimera( m_010, m_110, m_210, vvx, vx2);
- backwardInverseChimeraWithK(m_020, m_120, m_220, vvx, vx2, c3o1, c1o3);
- backwardChimera( m_001, m_101, m_201, vvx, vx2);
- backwardChimera( m_011, m_111, m_211, vvx, vx2);
- backwardChimera( m_021, m_121, m_221, vvx, vx2);
- backwardInverseChimeraWithK(m_002, m_102, m_202, vvx, vx2, c3o1, c1o3);
- backwardChimera( m_012, m_112, m_212, vvx, vx2);
- backwardInverseChimeraWithK(m_022, m_122, m_222, vvx, vx2, c9o1, c1o9);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // Y - Dir
- backwardInverseChimeraWithK(m_000, m_010, m_020, vvy, vy2, c6o1, c1o6);
- backwardChimera( m_001, m_011, m_021, vvy, vy2);
- backwardInverseChimeraWithK(m_002, m_012, m_022, vvy, vy2, c18o1, c1o18);
- backwardInverseChimeraWithK(m_100, m_110, m_120, vvy, vy2, c3o2, c2o3);
- backwardChimera( m_101, m_111, m_121, vvy, vy2);
- backwardInverseChimeraWithK(m_102, m_112, m_122, vvy, vy2, c9o2, c2o9);
- backwardInverseChimeraWithK(m_200, m_210, m_220, vvy, vy2, c6o1, c1o6);
- backwardChimera( m_201, m_211, m_221, vvy, vy2);
- backwardInverseChimeraWithK(m_202, m_212, m_222, vvy, vy2, c18o1, c1o18);
-
- ////////////////////////////////////////////////////////////////////////////////////
- // Z - Dir
- backwardInverseChimeraWithK(m_000, m_001, m_002, vvz, vz2, c36o1, c1o36);
- backwardInverseChimeraWithK(m_010, m_011, m_012, vvz, vz2, c9o1, c1o9);
- backwardInverseChimeraWithK(m_020, m_021, m_022, vvz, vz2, c36o1, c1o36);
- backwardInverseChimeraWithK(m_100, m_101, m_102, vvz, vz2, c9o1, c1o9);
- backwardInverseChimeraWithK(m_110, m_111, m_112, vvz, vz2, c9o4, c4o9);
- backwardInverseChimeraWithK(m_120, m_121, m_122, vvz, vz2, c9o1, c1o9);
- backwardInverseChimeraWithK(m_200, m_201, m_202, vvz, vz2, c36o1, c1o36);
- backwardInverseChimeraWithK(m_210, m_211, m_212, vvz, vz2, c9o1, c1o9);
- backwardInverseChimeraWithK(m_220, m_221, m_222, vvz, vz2, c36o1, c1o36);
-
- ////////////////////////////////////////////////////////////////////////////////////
- //! - Write distributions: style of reading and writing the distributions from/to
- //! stored arrays dependent on timestep is based on the esoteric twist algorithm
- //! <a href="https://doi.org/10.3390/computation5020019"><b>[ M. Geier et al. (2017),
- //! DOI:10.3390/computation5020019 ]</b></a>
- //!
- (dist.f[DIR_P00])[k_000] = f_M00;
- (dist.f[DIR_M00])[k_M00] = f_P00;
- (dist.f[DIR_0P0])[k_000] = f_0M0;
- (dist.f[DIR_0M0])[k_0M0] = f_0P0;
- (dist.f[DIR_00P])[k_000] = f_00M;
- (dist.f[DIR_00M])[k_00M] = f_00P;
- (dist.f[DIR_PP0])[k_000] = f_MM0;
- (dist.f[DIR_MM0])[k_MM0] = f_PP0;
- (dist.f[DIR_PM0])[k_0M0] = f_MP0;
- (dist.f[DIR_MP0])[k_M00] = f_PM0;
- (dist.f[DIR_P0P])[k_000] = f_M0M;
- (dist.f[DIR_M0M])[k_M0M] = f_P0P;
- (dist.f[DIR_P0M])[k_00M] = f_M0P;
- (dist.f[DIR_M0P])[k_M00] = f_P0M;
- (dist.f[DIR_0PP])[k_000] = f_0MM;
- (dist.f[DIR_0MM])[k_0MM] = f_0PP;
- (dist.f[DIR_0PM])[k_00M] = f_0MP;
- (dist.f[DIR_0MP])[k_0M0] = f_0PM;
- (dist.f[DIR_000])[k_000] = f_000;
- (dist.f[DIR_PPP])[k_000] = f_MMM;
- (dist.f[DIR_PMP])[k_0M0] = f_MPM;
- (dist.f[DIR_PPM])[k_00M] = f_MMP;
- (dist.f[DIR_PMM])[k_0MM] = f_MPP;
- (dist.f[DIR_MPP])[k_M00] = f_PMM;
- (dist.f[DIR_MMP])[k_MM0] = f_PPM;
- (dist.f[DIR_MPM])[k_M0M] = f_PMP;
- (dist.f[DIR_MMM])[k_MMM] = f_PPP;
- }
-}
-
-template __global__ void LB_Kernel_TurbulentViscosityCumulantK17CompChim < TurbulenceModel::AMD > ( real omega_in, uint* typeOfGridNode, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, bool bodyForce, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep);
-
-template __global__ void LB_Kernel_TurbulentViscosityCumulantK17CompChim < TurbulenceModel::Smagorinsky > ( real omega_in, uint* typeOfGridNode, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, bool bodyForce, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep);
-
-template __global__ void LB_Kernel_TurbulentViscosityCumulantK17CompChim < TurbulenceModel::QR > ( real omega_in, uint* typeOfGridNode, uint* neighborX, uint* neighborY, uint* neighborZ, real* distributions, real* rho, real* vx, real* vy, real* vz, real* turbulentViscosity, real SGSconstant, unsigned long size_Mat, int level, bool bodyForce, real* forces, real* bodyForceX, real* bodyForceY, real* bodyForceZ, real* quadricLimiters, bool isEvenTimestep);
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/TurbulentViscosityFluidFlowCompStrategy.cpp b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/TurbulentViscosityFluidFlowCompStrategy.cpp
deleted file mode 100644
index f3615a89994f0ca1fafdc1eda905d3c3b615d478..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/TurbulentViscosityFluidFlowCompStrategy.cpp
+++ /dev/null
@@ -1,23 +0,0 @@
-#include "TurbulentViscosityFluidFlowCompStrategy.h"
-
-#include "Parameter/Parameter.h"
-
-std::shared_ptr<TurbulentViscosityFluidFlowCompStrategy> TurbulentViscosityFluidFlowCompStrategy::getInstance()
-{
- static std::shared_ptr<TurbulentViscosityFluidFlowCompStrategy> uniqueInstance;
- if (!uniqueInstance)
- uniqueInstance = std::shared_ptr<TurbulentViscosityFluidFlowCompStrategy>(new TurbulentViscosityFluidFlowCompStrategy());
- return uniqueInstance;
-}
-
-bool TurbulentViscosityFluidFlowCompStrategy::checkParameter(std::shared_ptr<Parameter> para)
-{
- if (!para->getUseTurbulentViscosity())
- return false;
- else if (!para->getCompOn())
- return false;
- else
- return true;
-}
-
-TurbulentViscosityFluidFlowCompStrategy::TurbulentViscosityFluidFlowCompStrategy() {}
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/TurbulentViscosityFluidFlowCompStrategy.h b/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/TurbulentViscosityFluidFlowCompStrategy.h
deleted file mode 100644
index 95eff17777f7f0d1c3e05fe1b0d93892a88646a4..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/TurbulentViscosityFluidFlowCompStrategy.h
+++ /dev/null
@@ -1,18 +0,0 @@
-#ifndef AMD_FLUID_FLOW_COMP_STRATEGY_H
-#define AMD_FLUID_FLOW_COMP_STRATEGY_H
-
-#include "Kernel/Utilities/CheckParameterStrategy/CheckParameterStrategy.h"
-
-
-class TurbulentViscosityFluidFlowCompStrategy : public CheckParameterStrategy
-{
-public:
- static std::shared_ptr<TurbulentViscosityFluidFlowCompStrategy> getInstance();
-
- bool checkParameter(std::shared_ptr<Parameter> para);
-
-private:
- TurbulentViscosityFluidFlowCompStrategy();
-
-};
-#endif
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/Kernel/Utilities/KernelFactory/KernelFactoryImp.cpp b/src/gpu/VirtualFluids_GPU/Kernel/Utilities/KernelFactory/KernelFactoryImp.cpp
index 53ec240f096080097416e640fdd095c3812fb34c..5a2d8c9a426e5cb23ca75f91aaf6fbff75cba72b 100644
--- a/src/gpu/VirtualFluids_GPU/Kernel/Utilities/KernelFactory/KernelFactoryImp.cpp
+++ b/src/gpu/VirtualFluids_GPU/Kernel/Utilities/KernelFactory/KernelFactoryImp.cpp
@@ -8,11 +8,9 @@
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/BGKPlus/BGKPlusCompSP27.h"
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/Cascade/CascadeCompSP27.h"
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/Cumulant/CumulantCompSP27.h"
-#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17Comp.h"
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17Unified/CumulantK17Unified.h"
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chim/CumulantK17CompChim.h"
-#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimStream/CumulantK17CompChimStream.h"
-#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17chimRedesigned/CumulantK17CompChimRedesigned.h"
+#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17/CumulantK17.h"
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK17Bulk/CumulantK17BulkComp.h"
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantAll4/CumulantAll4CompSP27.h"
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/CumulantK18/CumulantK18Comp.h"
@@ -49,9 +47,6 @@
#include "Kernel/Kernels/WaleKernels/FluidFlow/Compressible/CumulantK15/WaleCumulantK15Comp.h"
#include "Kernel/Kernels/WaleKernels/FluidFlow/Compressible/CumulantK15BySoniMalav/WaleBySoniMalavCumulantK15Comp.h"
-//turbulent viscosity kernel
-#include "Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/CumulantK17chim/TurbulentViscosityCumulantK17CompChim.h"
-
//strategies
#include "Kernel/Kernels/BasicKernels/FluidFlow/Compressible/FluidFlowCompStrategy.h"
#include "Kernel/Kernels/BasicKernels/FluidFlow/Incompressible/FluidFlowIncompStrategy.h"
@@ -61,7 +56,6 @@
#include "Kernel/Kernels/BasicKernels/AdvectionDiffusion/Incompressible/Mod7/ADMod7IncompStrategy.h"
#include "Kernel/Kernels/PorousMediaKernels/FluidFlow/Compressible/PMFluidFlowCompStrategy.h"
#include "Kernel/Kernels/WaleKernels/FluidFlow/Compressible/WaleFluidFlowCompStrategy.h"
-#include "Kernel/Kernels/TurbulentViscosityKernels/FluidFlow/Compressible/TurbulentViscosityFluidFlowCompStrategy.h"
std::vector<std::shared_ptr<Kernel>> KernelFactoryImp::makeKernels(std::shared_ptr<Parameter> para)
{
@@ -118,9 +112,6 @@ std::shared_ptr<Kernel> KernelFactoryImp::makeKernel(std::shared_ptr<Parameter>
} else if (kernel == "CumulantCompSP27") {
newKernel = CumulantCompSP27::getNewInstance(para, level);
checkStrategy = FluidFlowCompStrategy::getInstance();
- } else if (kernel == "CumulantK17Comp") {
- newKernel = CumulantK17Comp::getNewInstance(para, level);
- checkStrategy = FluidFlowCompStrategy::getInstance();
} else if (kernel == "CumulantK15Unified") {
newKernel = std::make_shared<vf::gpu::CumulantK15Unified>(para, level);
checkStrategy = FluidFlowCompStrategy::getInstance();
@@ -133,12 +124,26 @@ std::shared_ptr<Kernel> KernelFactoryImp::makeKernel(std::shared_ptr<Parameter>
} else if (kernel == "CumulantK17CompChim") {
newKernel = CumulantK17CompChim::getNewInstance(para, level);
checkStrategy = FluidFlowCompStrategy::getInstance();
- } else if (kernel == "CumulantK17CompChimStream") {
- newKernel = CumulantK17CompChimStream::getNewInstance(para, level);
- checkStrategy = FluidFlowCompStrategy::getInstance();
- } else if (kernel == "CumulantK17CompChimRedesigned") {
- newKernel = CumulantK17CompChimRedesigned::getNewInstance(para, level);
- checkStrategy = FluidFlowCompStrategy::getInstance();
+ } else if (kernel == "CumulantK17"){
+ switch(para->getTurbulenceModel())
+ {
+ case TurbulenceModel::AMD:
+ newKernel = CumulantK17<TurbulenceModel::AMD>::getNewInstance(para, level);
+ break;
+ case TurbulenceModel::Smagorinsky:
+ newKernel = CumulantK17<TurbulenceModel::Smagorinsky>::getNewInstance(para, level);
+ break;
+ case TurbulenceModel::QR:
+ newKernel = CumulantK17<TurbulenceModel::QR>::getNewInstance(para, level);
+ break;
+ case TurbulenceModel::None:
+ newKernel = CumulantK17<TurbulenceModel::None>::getNewInstance(para, level);
+ break;
+ default:
+ throw std::runtime_error("Unknown turbulence model!");
+ break;
+ }
+ checkStrategy = FluidFlowCompStrategy::getInstance();
} else if (kernel == "CumulantAll4CompSP27") {
newKernel = CumulantAll4CompSP27::getNewInstance(para, level);
checkStrategy = FluidFlowCompStrategy::getInstance();
@@ -197,35 +202,9 @@ std::shared_ptr<Kernel> KernelFactoryImp::makeKernel(std::shared_ptr<Parameter>
newKernel = WaleBySoniMalavCumulantK15Comp::getNewInstance(para, level);// ||
checkStrategy = WaleFluidFlowCompStrategy::getInstance(); // wale model
} //===============
- else if (kernel == "TurbulentViscosityCumulantK17CompChim"){ // compressible with turbulent viscosity
- switch(para->getTurbulenceModel()) // ||
- { // \/ //
- case TurbulenceModel::AMD:
- newKernel = TurbulentViscosityCumulantK17CompChim<TurbulenceModel::AMD>::getNewInstance(para, level);
- break;
- case TurbulenceModel::Smagorinsky:
- newKernel = TurbulentViscosityCumulantK17CompChim<TurbulenceModel::Smagorinsky>::getNewInstance(para, level);
- break;
- case TurbulenceModel::QR:
- newKernel = TurbulentViscosityCumulantK17CompChim<TurbulenceModel::QR>::getNewInstance(para, level);
- break;
- case TurbulenceModel::None:
- throw std::runtime_error("TurbulentViscosityCumulantK17CompChim currently not implemented for TurbulenceModel::None!");
- break;
- default:
- throw std::runtime_error("Unknown turbulence model!");
- break;
- }
- checkStrategy = TurbulentViscosityFluidFlowCompStrategy::getInstance();
- // /\ //
- // ||
- // compressible with turbulent viscosity
- //===============
- }
else {
throw std::runtime_error("KernelFactory does not know the KernelType.");
}
-
newKernel->setCheckParameterStrategy(checkStrategy);
para->setKernelNeedsFluidNodeIndicesToRun(newKernel->getKernelUsesFluidNodeIndices());
return newKernel;
diff --git a/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManager.cpp b/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManager.cpp
index cc945ea225a28c58dca4ceefdb80fffb76228b21..1310cf09e5ce05270e88cd3a53e4910816bd7628 100644
--- a/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManager.cpp
+++ b/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManager.cpp
@@ -38,6 +38,7 @@
#include "BCKernelManager.h"
#include "Factories/BoundaryConditionFactory.h"
+#include "GridGenerator/TransientBCSetter/TransientBCSetter.h"
#include "Calculation/Cp.h"
#include "Calculation/DragLift.h"
#include "GPU/GPU_Interface.h"
@@ -51,6 +52,7 @@ BCKernelManager::BCKernelManager(SPtr<Parameter> parameter, BoundaryConditionFac
this->pressureBoundaryConditionPre = bcFactory->getPressureBoundaryConditionPre();
this->geometryBoundaryConditionPost = bcFactory->getGeometryBoundaryConditionPost();
this->stressBoundaryConditionPost = bcFactory->getStressBoundaryConditionPost();
+ this->precursorBoundaryConditionPost = bcFactory->getPrecursorBoundaryConditionPost();
checkBoundaryCondition(this->velocityBoundaryConditionPost, this->para->getParD(0)->velocityBC,
"velocityBoundaryConditionPost");
@@ -64,6 +66,8 @@ BCKernelManager::BCKernelManager(SPtr<Parameter> parameter, BoundaryConditionFac
"geometryBoundaryConditionPost");
checkBoundaryCondition(this->stressBoundaryConditionPost, this->para->getParD(0)->stressBC,
"stressBoundaryConditionPost");
+ checkBoundaryCondition(this->precursorBoundaryConditionPost, this->para->getParD(0)->precursorBC,
+ "precursorBoundaryConditionPost");
}
void BCKernelManager::runVelocityBCKernelPre(const int level) const
@@ -387,3 +391,41 @@ void BCKernelManager::runNoSlipBCKernelPost(const int level) const{
noSlipBoundaryConditionPost(para->getParD(level).get(), &(para->getParD(level)->noSlipBC));
}
}
+
+void BCKernelManager::runPrecursorBCKernelPost(int level, uint t, CudaMemoryManager* cudaMemoryManager)
+{
+ if(para->getParH(level)->precursorBC.numberOfBCnodes == 0) return;
+
+ uint t_level = para->getTimeStep(level, t, true);
+
+ uint lastTime = (para->getParD(level)->precursorBC.nPrecursorReads-2)*para->getParD(level)->precursorBC.timeStepsBetweenReads; // timestep currently loaded into last arrays
+ uint currentTime = (para->getParD(level)->precursorBC.nPrecursorReads-1)*para->getParD(level)->precursorBC.timeStepsBetweenReads; // timestep currently loaded into current arrays
+ uint nextTime = para->getParD(level)->precursorBC.nPrecursorReads *para->getParD(level)->precursorBC.timeStepsBetweenReads; // timestep currently loaded into next arrays
+
+ if(t_level>=currentTime)
+ {
+ //cycle time
+ lastTime = currentTime;
+ currentTime = nextTime;
+ nextTime += para->getParD(level)->precursorBC.timeStepsBetweenReads;
+
+ //cycle pointers
+ real* tmp = para->getParD(level)->precursorBC.last;
+ para->getParD(level)->precursorBC.last = para->getParD(level)->precursorBC.current;
+ para->getParD(level)->precursorBC.current = para->getParD(level)->precursorBC.next;
+ para->getParD(level)->precursorBC.next = tmp;
+
+ real loadTime = nextTime*pow(2,-level)*para->getTimeRatio();
+
+ for(auto reader : para->getParH(level)->transientBCInputFileReader)
+ {
+ reader->getNextData(para->getParH(level)->precursorBC.next, para->getParH(level)->precursorBC.numberOfPrecursorNodes, loadTime);
+ }
+ cudaMemoryManager->cudaCopyPrecursorData(level);
+ para->getParD(level)->precursorBC.nPrecursorReads++;
+ para->getParH(level)->precursorBC.nPrecursorReads++;
+ }
+
+ real tRatio = real(t_level-lastTime)/para->getParD(level)->precursorBC.timeStepsBetweenReads;
+ precursorBoundaryConditionPost(para->getParD(level).get(), ¶->getParD(level)->precursorBC, tRatio, para->getVelocityRatio());
+}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManager.h b/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManager.h
index 423a9cc9056281a3a2a135ae32fa26cc47f93967..339100e6b5307e8e60f8d0846560bf89c6eea1a1 100644
--- a/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManager.h
+++ b/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManager.h
@@ -41,6 +41,7 @@
#include "PointerDefinitions.h"
#include "VirtualFluids_GPU_export.h"
+
class CudaMemoryManager;
class BoundaryConditionFactory;
class Parameter;
@@ -48,6 +49,7 @@ struct LBMSimulationParameter;
using boundaryCondition = std::function<void(LBMSimulationParameter *, QforBoundaryConditions *)>;
using boundaryConditionWithParameter = std::function<void(Parameter *, QforBoundaryConditions *, const int level)>;
+using precursorBoundaryCondition = std::function<void(LBMSimulationParameter *, QforPrecursorBoundaryConditions *, real tRatio, real velocityRatio)>;
//! \class BCKernelManager
//! \brief manage the cuda kernel calls to boundary conditions
@@ -84,7 +86,10 @@ public:
//! \brief calls the device function of the pressure boundary condition (post-collision)
void runPressureBCKernelPost(const int level) const;
- //! \brief calls the device function of the outflow boundary condition (pre-collision)
+ //! \brief calls the device function of the precursor boundary condition
+ void runPrecursorBCKernelPost(int level, uint t, CudaMemoryManager* cudaMemoryManager);
+
+ //! \brief calls the device function of the outflow boundary condition
void runOutflowBCKernelPre(const int level) const;
//! \brief calls the device function of the stress wall model (post-collision)
@@ -96,13 +101,16 @@ private:
//! \param boundaryCondition: a kernel function for the boundary condition
//! \param bcStruct: a struct containing the grid nodes which are part of the boundary condition
//! \param bcName: the name of the checked boundary condition
- template <typename bcFunction>
- void checkBoundaryCondition(const bcFunction &boundaryCondition, const QforBoundaryConditions &bcStruct, const std::string &bcName)
+ template <typename bcFunction, typename QforBC>
+ void checkBoundaryCondition(const bcFunction &boundaryCondition, const QforBC &bcStruct, const std::string &bcName)
{
if (!boundaryCondition && bcStruct.numberOfBCnodes > 0)
throw std::runtime_error("The boundary condition " + bcName + " was not set!");
}
+ void runDistributionPrecursorBCKernelPost(int level, uint t, CudaMemoryManager* cudaMemoryManager);
+ void runVelocityPrecursorBCKernelPost(int level, uint t, CudaMemoryManager* cudaMemoryManager);
+
SPtr<Parameter> para;
boundaryCondition velocityBoundaryConditionPost = nullptr;
@@ -111,5 +119,6 @@ private:
boundaryCondition pressureBoundaryConditionPre = nullptr;
boundaryCondition geometryBoundaryConditionPost = nullptr;
boundaryConditionWithParameter stressBoundaryConditionPost = nullptr;
+ precursorBoundaryCondition precursorBoundaryConditionPost = nullptr;
};
#endif
diff --git a/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManagerTest.cpp b/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManagerTest.cpp
index d55fa51bd8a225dd4e89e684bc81cd56f3f450c0..a0e02112e821eedcfeb013d3465529f668309529 100644
--- a/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManagerTest.cpp
+++ b/src/gpu/VirtualFluids_GPU/KernelManager/BCKernelManagerTest.cpp
@@ -53,3 +53,9 @@ TEST_F(BCKernelManagerTest_BCsNotSpecified, stressBoundaryConditionPost_NotSpeci
para->getParD(0)->stressBC.numberOfBCnodes = 1;
EXPECT_THROW(BCKernelManager(para, &bcFactory), std::runtime_error);
}
+
+TEST_F(BCKernelManagerTest_BCsNotSpecified, precursorBoundaryConditionPost_NotSpecified)
+{
+ para->getParD(0)->precursorBC.numberOfBCnodes = 1;
+ EXPECT_THROW(BCKernelManager(para, &bcFactory), std::runtime_error);
+}
diff --git a/src/gpu/VirtualFluids_GPU/KernelManager/GridScalingKernelManager.cpp b/src/gpu/VirtualFluids_GPU/KernelManager/GridScalingKernelManager.cpp
index c3129e31a9c750a012a26d58961062eaf3f40add..2b6a266c0d4e5f523091fa4982eee5d83b2ec675 100644
--- a/src/gpu/VirtualFluids_GPU/KernelManager/GridScalingKernelManager.cpp
+++ b/src/gpu/VirtualFluids_GPU/KernelManager/GridScalingKernelManager.cpp
@@ -59,8 +59,9 @@ GridScalingKernelManager::GridScalingKernelManager(SPtr<Parameter> parameter, Gr
VF_LOG_TRACE("Function for scalingCoarseToFine is nullptr");
}
-void GridScalingKernelManager::runFineToCoarseKernelLB(const int level, InterpolationCellFC *icellFC, OffFC &offFC, int streamIndex) const{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
+void GridScalingKernelManager::runFineToCoarseKernelLB(const int level, InterpolationCellFC *icellFC, OffFC &offFC, CudaStreamIndex streamIndex) const
+{
+ cudaStream_t stream = para->getStreamManager()->getStream(streamIndex);
this->scalingFineToCoarse(para->getParD(level).get(), para->getParD(level+1).get(), icellFC, offFC, stream);
@@ -327,9 +328,9 @@ void GridScalingKernelManager::runFineToCoarseKernelAD(const int level) const
}
}
-void GridScalingKernelManager::runCoarseToFineKernelLB(const int level, InterpolationCellCF* icellCF, OffCF &offCF, int streamIndex) const
+void GridScalingKernelManager::runCoarseToFineKernelLB(const int level, InterpolationCellCF* icellCF, OffCF &offCF, CudaStreamIndex streamIndex) const
{
- cudaStream_t stream = (streamIndex == -1) ? CU_STREAM_LEGACY : para->getStreamManager()->getStream(streamIndex);
+ cudaStream_t stream = para->getStreamManager()->getStream(streamIndex);
this->scalingCoarseToFine(para->getParD(level).get(), para->getParD(level+1).get(), icellCF, offCF, stream);
// ScaleCF_comp_D3Q27F3(
diff --git a/src/gpu/VirtualFluids_GPU/KernelManager/GridScalingKernelManager.h b/src/gpu/VirtualFluids_GPU/KernelManager/GridScalingKernelManager.h
index 85cdd88ec2e3a6622108026ce8f53c5c770f8afe..3c78ee7f9db254556e8ec6dbbafaf51cd995f10b 100644
--- a/src/gpu/VirtualFluids_GPU/KernelManager/GridScalingKernelManager.h
+++ b/src/gpu/VirtualFluids_GPU/KernelManager/GridScalingKernelManager.h
@@ -44,6 +44,7 @@
class Parameter;
class CudaMemoryManager;
class GridScalingFactory;
+enum class CudaStreamIndex;
struct LBMSimulationParameter;
struct CUstream_st;
@@ -62,14 +63,14 @@ public:
//! \throws std::runtime_error when the user forgets to specify a scaling function
GridScalingKernelManager(SPtr<Parameter> parameter, GridScalingFactory *gridScalingFactory);
- //! \brief calls the device function of the fine to coarse grid interpolation kernel
- void runFineToCoarseKernelLB(const int level, InterpolationCellFC *icellFC, OffFC &offFC, int streamIndex) const;
+ //! \brief calls the device function of the fine to coarse grid interpolation kernelH
+ void runFineToCoarseKernelLB(const int level, InterpolationCellFC *icellFC, OffFC &offFC, CudaStreamIndex streamIndex) const;
//! \brief calls the device function of the fine to coarse grid interpolation kernel (advection diffusion)
void runFineToCoarseKernelAD(const int level) const;
//! \brief calls the device function of the coarse to fine grid interpolation kernel
- void runCoarseToFineKernelLB(const int level, InterpolationCellCF *icellCF, OffCF &offCF, int streamIndex) const;
+ void runCoarseToFineKernelLB(const int level, InterpolationCellCF *icellCF, OffCF &offCF, CudaStreamIndex streamIndex) const;
//! \brief calls the device function of the coarse to fine grid interpolation kernel (advection diffusion)
void runCoarseToFineKernelAD(const int level) const;
diff --git a/src/gpu/VirtualFluids_GPU/LBM/LB.h b/src/gpu/VirtualFluids_GPU/LBM/LB.h
index eea4adfda3c1ef0862f39ef58fc6e065af7bab1b..a7517ca0f07b9c1379501aa3ff750637d230a7bd 100644
--- a/src/gpu/VirtualFluids_GPU/LBM/LB.h
+++ b/src/gpu/VirtualFluids_GPU/LBM/LB.h
@@ -53,17 +53,33 @@
//! \brief An enumeration for selecting a turbulence model
enum class TurbulenceModel {
//! - Smagorinsky
- Smagorinsky,
+ Smagorinsky,
//! - AMD (Anisotropic Minimum Dissipation) model, see e.g. Rozema et al., Phys. Fluids 27, 085107 (2015), https://doi.org/10.1063/1.4928700
- AMD,
+ AMD,
//! - QR model by Verstappen
- QR,
+ QR,
//! - TODO: move the WALE model here from the old kernels
//WALE
//! - No turbulence model
- None
+ None
};
+//! \brief An enumeration for selecting a template of the collision kernel (CumulantK17)
+enum class CollisionTemplate {
+ //! - Default: plain collision without additional read/write
+ Default,
+ //! - WriteMacroVars: collision \w write out macroscopic variables
+ WriteMacroVars,
+ //! - ApplyBodyForce: collision \w read and apply body force in the collision kernel
+ ApplyBodyForce,
+ //! - AllFeatures: collision \w write out macroscopic variables AND read and apply body force
+ AllFeatures,
+ //! - Border: collision on border nodes
+ SubDomainBorder
+};
+constexpr std::initializer_list<CollisionTemplate> all_CollisionTemplate = { CollisionTemplate::Default, CollisionTemplate::WriteMacroVars, CollisionTemplate::ApplyBodyForce, CollisionTemplate::AllFeatures, CollisionTemplate::SubDomainBorder};
+constexpr std::initializer_list<CollisionTemplate> bulk_CollisionTemplate = { CollisionTemplate::Default, CollisionTemplate::WriteMacroVars, CollisionTemplate::ApplyBodyForce, CollisionTemplate::AllFeatures};
+
struct InitCondition
{
real Re;
@@ -144,6 +160,7 @@ struct InitCondition
bool hasWallModelMonitor {false};
bool simulatePorousMedia {false};
bool streetVelocityFile {false};
+ real outflowPressureCorrectionFactor {0.0};
};
//Interface Cells
@@ -214,6 +231,21 @@ typedef struct QforBC{
real *normalX, *normalY, *normalZ;
}QforBoundaryConditions;
+typedef struct QforPrecursorBC{
+ int* k;
+ int numberOfBCnodes=0;
+ int sizeQ;
+ int numberOfPrecursorNodes=0;
+ uint nPrecursorReads=0;
+ uint timeStepsBetweenReads;
+ size_t numberOfQuantities;
+ real* q27[27];
+ uint* planeNeighbor0PP, *planeNeighbor0PM, *planeNeighbor0MP, *planeNeighbor0MM;
+ real* weights0PP, *weights0PM, *weights0MP, *weights0MM;
+ real* last, *current, *next;
+ real velocityX, velocityY, velocityZ;
+}QforPrecursorBoundaryConditions;
+
//BCTemp
typedef struct TempforBC{
int* k;
@@ -249,6 +281,7 @@ typedef struct WMparas{
real* Fz;
}WallModelParameters;
+
//measurePoints
typedef struct MeasP{
std::string name;
diff --git a/src/gpu/VirtualFluids_GPU/LBM/Simulation.cpp b/src/gpu/VirtualFluids_GPU/LBM/Simulation.cpp
index 703e935e6edb5676c7d6e045a38e3ec20d7a4b41..84ab84ff93fa7706bcc27d7e61a18f580f3c8dbe 100644
--- a/src/gpu/VirtualFluids_GPU/LBM/Simulation.cpp
+++ b/src/gpu/VirtualFluids_GPU/LBM/Simulation.cpp
@@ -97,11 +97,7 @@ void Simulation::init(GridProvider &gridProvider, BoundaryConditionFactory *bcFa
gridProvider.allocAndCopyForcing();
gridProvider.allocAndCopyQuadricLimiters();
- if (para->getKernelNeedsFluidNodeIndicesToRun()) {
- gridProvider.allocArrays_fluidNodeIndices();
- gridProvider.allocArrays_fluidNodeIndicesBorder();
- }
-
+
gridProvider.setDimensions();
gridProvider.setBoundingBox();
@@ -113,12 +109,7 @@ void Simulation::init(GridProvider &gridProvider, BoundaryConditionFactory *bcFa
para->setStartTurn((unsigned int)0); // 100000
restart_object = std::make_shared<ASCIIRestartObject>();
- //////////////////////////////////////////////////////////////////////////
- // CUDA streams
- if (para->getUseStreams()) {
- para->getStreamManager()->launchStreams(2u);
- para->getStreamManager()->createCudaEvents();
- }
+
//////////////////////////////////////////////////////////////////////////
VF_LOG_INFO("LB_Modell: D3Q{}", para->getD3Qxx());
VF_LOG_INFO("Re: {}", para->getRe());
@@ -134,14 +125,32 @@ void Simulation::init(GridProvider &gridProvider, BoundaryConditionFactory *bcFa
//////////////////////////////////////////////////////////////////////////
allocNeighborsOffsetsScalesAndBoundaries(gridProvider);
+ //! Get tagged fluid nodes with corresponding value for CollisionTemplate from interactors
for (SPtr<PreCollisionInteractor> actuator : para->getActuators()) {
actuator->init(para.get(), &gridProvider, cudaMemoryManager.get());
+ actuator->getTaggedFluidNodes( para.get(), &gridProvider );
}
for (SPtr<PreCollisionInteractor> probe : para->getProbes()) {
probe->init(para.get(), &gridProvider, cudaMemoryManager.get());
+ probe->getTaggedFluidNodes( para.get(), &gridProvider );
}
+ //////////////////////////////////////////////////////////////////////////
+ // CUDA streams
+ if (para->getUseStreams()) {
+ para->getStreamManager()->registerStream(CudaStreamIndex::SubDomainBorder);
+ para->getStreamManager()->registerStream(CudaStreamIndex::Bulk);
+ para->getStreamManager()->launchStreams();
+ para->getStreamManager()->createCudaEvents();
+ }
+ //////////////////////////////////////////////////////////////////////////
+
+ if (para->getKernelNeedsFluidNodeIndicesToRun())
+ {
+ gridProvider.sortFluidNodeTags();
+ gridProvider.allocArrays_taggedFluidNodes();
+ }
//////////////////////////////////////////////////////////////////////////
// Kernel init
//////////////////////////////////////////////////////////////////////////
diff --git a/src/gpu/VirtualFluids_GPU/Output/DistributionDebugInspector.cu b/src/gpu/VirtualFluids_GPU/Output/DistributionDebugInspector.cu
new file mode 100644
index 0000000000000000000000000000000000000000..9ff86ea36c5b3c465990d622547d98cb4686f929
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/Output/DistributionDebugInspector.cu
@@ -0,0 +1,143 @@
+
+#include "DistributionDebugInspector.h"
+
+#include "Parameter/Parameter.h"
+#include "LBM/LB.h"
+#include "lbm/constants/D3Q27.h"
+#include <lbm/constants/NumericConstants.h>
+#include "Kernel/Utilities/DistributionHelper.cuh"
+
+#include <cuda/CudaGrid.h>
+#include <cuda.h>
+
+#include <iostream>
+
+using namespace vf::lbm::constant;
+using namespace vf::lbm::dir;
+
+__global__ void printFs( real* distributions,
+ bool isEvenTimestep,
+ unsigned long numberOfFluidNodes,
+ uint* neighborX,
+ uint* neighborY,
+ uint* neighborZ,
+ uint* typeOfGridNode,
+ real* coordX,
+ real* coordY,
+ real* coordZ,
+ real minX,
+ real maxX,
+ real minY,
+ real maxY,
+ real minZ,
+ real maxZ)
+ {
+ const unsigned k_000 = vf::gpu::getNodeIndex();
+
+ if (k_000 >= numberOfFluidNodes || typeOfGridNode[k_000]!=GEO_FLUID )
+ return;
+
+ real coordNodeX = coordX[k_000];
+ real coordNodeY = coordY[k_000];
+ real coordNodeZ = coordZ[k_000];
+
+ if( coordNodeX>=minX && coordNodeX<=maxX &&
+ coordNodeY>=minY && coordNodeY<=maxY &&
+ coordNodeZ>=minZ && coordNodeZ<=maxZ )
+ {
+ Distributions27 dist = vf::gpu::getDistributionReferences27(distributions, numberOfFluidNodes, isEvenTimestep);
+ ////////////////////////////////////////////////////////////////////////////////
+ //! - Set neighbor indices (necessary for indirect addressing)
+ uint k_M00 = neighborX[k_000];
+ uint k_0M0 = neighborY[k_000];
+ uint k_00M = neighborZ[k_000];
+ uint k_MM0 = neighborY[k_M00];
+ uint k_M0M = neighborZ[k_M00];
+ uint k_0MM = neighborZ[k_0M0];
+ uint k_MMM = neighborZ[k_MM0];
+ ////////////////////////////////////////////////////////////////////////////////////
+ //! - Set local distributions
+ //!
+ real f_000 = (dist.f[DIR_000])[k_000];
+ real f_P00 = (dist.f[DIR_P00])[k_000];
+ real f_M00 = (dist.f[DIR_M00])[k_M00];
+ real f_0P0 = (dist.f[DIR_0P0])[k_000];
+ real f_0M0 = (dist.f[DIR_0M0])[k_0M0];
+ real f_00P = (dist.f[DIR_00P])[k_000];
+ real f_00M = (dist.f[DIR_00M])[k_00M];
+ real f_PP0 = (dist.f[DIR_PP0])[k_000];
+ real f_MM0 = (dist.f[DIR_MM0])[k_MM0];
+ real f_PM0 = (dist.f[DIR_PM0])[k_0M0];
+ real f_MP0 = (dist.f[DIR_MP0])[k_M00];
+ real f_P0P = (dist.f[DIR_P0P])[k_000];
+ real f_M0M = (dist.f[DIR_M0M])[k_M0M];
+ real f_P0M = (dist.f[DIR_P0M])[k_00M];
+ real f_M0P = (dist.f[DIR_M0P])[k_M00];
+ real f_0PP = (dist.f[DIR_0PP])[k_000];
+ real f_0MM = (dist.f[DIR_0MM])[k_0MM];
+ real f_0PM = (dist.f[DIR_0PM])[k_00M];
+ real f_0MP = (dist.f[DIR_0MP])[k_0M0];
+ real f_PPP = (dist.f[DIR_PPP])[k_000];
+ real f_MPP = (dist.f[DIR_MPP])[k_M00];
+ real f_PMP = (dist.f[DIR_PMP])[k_0M0];
+ real f_MMP = (dist.f[DIR_MMP])[k_MM0];
+ real f_PPM = (dist.f[DIR_PPM])[k_00M];
+ real f_MPM = (dist.f[DIR_MPM])[k_M0M];
+ real f_PMM = (dist.f[DIR_PMM])[k_0MM];
+ real f_MMM = (dist.f[DIR_MMM])[k_MMM];
+
+ real drho = ((((f_PPP + f_MMM) + (f_MPM + f_PMP)) + ((f_MPP + f_PMM) + (f_MMP + f_PPM))) +
+ (((f_0MP + f_0PM) + (f_0MM + f_0PP)) + ((f_M0P + f_P0M) + (f_M0M + f_P0P)) +
+ ((f_MP0 + f_PM0) + (f_MM0 + f_PP0))) +
+ ((f_M00 + f_P00) + (f_0M0 + f_0P0) + (f_00M + f_00P))) +
+ f_000;
+
+ real oneOverRho = c1o1 / (c1o1 + drho);
+
+ real vvx = ((((f_PPP - f_MMM) + (f_PMP - f_MPM)) + ((f_PMM - f_MPP) + (f_PPM - f_MMP))) +
+ (((f_P0M - f_M0P) + (f_P0P - f_M0M)) + ((f_PM0 - f_MP0) + (f_PP0 - f_MM0))) + (f_P00 - f_M00)) *
+ oneOverRho;
+ real vvy = ((((f_PPP - f_MMM) + (f_MPM - f_PMP)) + ((f_MPP - f_PMM) + (f_PPM - f_MMP))) +
+ (((f_0PM - f_0MP) + (f_0PP - f_0MM)) + ((f_MP0 - f_PM0) + (f_PP0 - f_MM0))) + (f_0P0 - f_0M0)) *
+ oneOverRho;
+ real vvz = ((((f_PPP - f_MMM) + (f_PMP - f_MPM)) + ((f_MPP - f_PMM) + (f_MMP - f_PPM))) +
+ (((f_0MP - f_0PM) + (f_0PP - f_0MM)) + ((f_M0P - f_P0M) + (f_P0P - f_M0M))) + (f_00P - f_00M)) *
+ oneOverRho;
+
+ printf("Node %u \t (%f\t%f\t%f)\n rho: %f\t velo: %f\t %f \t %f\n\n" , k_000, coordNodeX, coordNodeY, coordNodeZ, drho, vvx, vvy, vvz);
+ printf("Node %u \t (%f\t%f\t%f)\n f_M00\t%f\t f_000\t%f\t f_P00\t%f\n f_MP0\t%f\t f_0P0\t%f\t f_PP0\t%f\n f_MM0\t%f\t f_0M0\t%f\t f_PM0\t%f\n f_M0P\t%f\t f_00P\t%f\t f_P0P\t%f\n f_M0M\t%f\t f_00M\t%f\t f_P0M\t%f\n f_MPP\t%f\t f_0PP\t%f\t f_PPP\t%f\n f_MPM\t%f\t f_0PM\t%f\t f_PPM\t%f\n f_MMP\t%f\t f_0MP\t%f\t f_PMP\t%f\n f_MMM\t%f\t f_0MM\t%f\t f_PMM\t%f\n\n\n" , k_000, coordNodeX, coordNodeY, coordNodeZ, f_M00, f_000, f_P00,f_MP0, f_0P0, f_PP0, f_MM0, f_0M0, f_PM0, f_M0P, f_00P, f_P0P, f_M0M, f_00M, f_P0M, f_MPP, f_0PP, f_PPP, f_MPM, f_0PM, f_PPM, f_MMP, f_0MP, f_PMP, f_MMM, f_0MM, f_PMM);
+
+ }
+
+ }
+
+
+
+
+void DistributionDebugInspector::inspect(std::shared_ptr<Parameter> para, uint level, uint t){
+
+ if(this->inspectionLevel!=level)
+ return;
+
+ std::cout << tag << ": distributions on level " << level << " at t " << t << std::endl;
+
+ vf::cuda::CudaGrid cudaGrid = vf::cuda::CudaGrid(para->getParD(level)->numberofthreads, para->getParD(level)->numberOfNodes);
+
+ printFs <<< cudaGrid.grid, cudaGrid.threads >>>( para->getParD(level)->distributions.f[0],
+ para->getParD(level)->isEvenTimestep,
+ (unsigned long)para->getParD(level)->numberOfNodes,
+ para->getParD(level)->neighborX,
+ para->getParD(level)->neighborY,
+ para->getParD(level)->neighborZ,
+ para->getParD(level)->typeOfGridNode,
+ para->getParD(level)->coordinateX,
+ para->getParD(level)->coordinateY,
+ para->getParD(level)->coordinateZ,
+ minX,
+ maxX,
+ minY,
+ maxY,
+ minZ,
+ maxZ);
+
+}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/Output/DistributionDebugInspector.h b/src/gpu/VirtualFluids_GPU/Output/DistributionDebugInspector.h
new file mode 100644
index 0000000000000000000000000000000000000000..53e030c0d16116a4edef0135f9ab435c853fd66b
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/Output/DistributionDebugInspector.h
@@ -0,0 +1,76 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file DistributionDebugInspector.h
+//! \author Henrik Asmuth
+//! \date 13/012/2022
+//! \brief Basic debugging class to print out f's in a certain area of the domain
+//!
+//! Basic debugging class. Needs to be directly added in UpdateGrid (could potentially also be added as a proper Probe in the future)
+//! How to use: Define a part of the domain via min/max x, y, and z. The DistributionDebugInspector will print out all f's in that area.
+//!
+//=======================================================================================
+
+#ifndef DISTRIBUTION_INSPECTOR_H
+#define DISTRIBUTION_INSPECTOR_H
+
+#include "Parameter/Parameter.h"
+
+
+class DistributionDebugInspector
+{
+public:
+ DistributionDebugInspector(uint _inspectionLevel, real _minX, real _maxX, real _minY, real _maxY, real _minZ, real _maxZ, std::string _tag):
+ inspectionLevel(_inspectionLevel),
+ minX(_minX),
+ maxX(_maxX),
+ minY(_minY),
+ maxY(_maxY),
+ minZ(_minZ),
+ maxZ(_maxZ),
+ tag(_tag)
+ {};
+
+ ~DistributionDebugInspector(){}
+
+ void inspect(std::shared_ptr<Parameter> para, uint level, uint t);
+
+
+private:
+uint inspectionLevel;
+real minX;
+real maxX;
+real minY;
+real maxY;
+real minZ;
+real maxZ;
+std::string tag;
+
+};
+
+#endif
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/Parameter/CudaStreamManager.cpp b/src/gpu/VirtualFluids_GPU/Parameter/CudaStreamManager.cpp
index 3731836f336d91c1bc4cc5f1a8f5ea0a10bee0a6..3cc771e413134e90b0d09d8eeb6dfee791f8a1e2 100644
--- a/src/gpu/VirtualFluids_GPU/Parameter/CudaStreamManager.cpp
+++ b/src/gpu/VirtualFluids_GPU/Parameter/CudaStreamManager.cpp
@@ -31,25 +31,39 @@
#include <helper_cuda.h>
#include <iostream>
-void CudaStreamManager::launchStreams(uint numberOfStreams)
+void CudaStreamManager::registerStream(CudaStreamIndex streamIndex)
+{
+ if(streamIndex != CudaStreamIndex::Legacy)
+ cudaStreams.emplace(streamIndex, nullptr);
+}
+void CudaStreamManager::launchStreams()
{
- cudaStreams.resize(numberOfStreams);
- for (cudaStream_t &stream : cudaStreams)
- cudaStreamCreate(&stream);
+ for (auto &stream : cudaStreams)
+ cudaStreamCreate(&stream.second);
}
void CudaStreamManager::terminateStreams()
{
- for (cudaStream_t &stream : cudaStreams)
- cudaStreamDestroy(stream);
+ for (auto &stream : cudaStreams)
+ cudaStreamDestroy(stream.second);
}
-cudaStream_t &CudaStreamManager::getStream(uint streamIndex)
-{ return cudaStreams[streamIndex]; }
-
-int CudaStreamManager::getBorderStreamIndex() { return borderStreamIndex; }
+cudaStream_t &CudaStreamManager::getStream(CudaStreamIndex streamIndex, uint multiStreamIndex)
+{
+ if(streamIndex == CudaStreamIndex::Legacy) return legacyStream;
+ if(streamIsRegistered(streamIndex))
+ {
+ auto it = cudaStreams.find(streamIndex);
+ for(uint idx=0; idx<multiStreamIndex; idx++) it++;
+ return it->second;
+ }
+ return legacyStream;
+}
-int CudaStreamManager::getBulkStreamIndex() { return bulkStreamIndex; }
+bool CudaStreamManager::streamIsRegistered(CudaStreamIndex streamIndex)
+{
+ return cudaStreams.count(streamIndex) > 0;
+}
void CudaStreamManager::createCudaEvents()
{
@@ -61,12 +75,12 @@ void CudaStreamManager::destroyCudaEvents()
checkCudaErrors(cudaEventDestroy(startBulkKernel));
}
-void CudaStreamManager::triggerStartBulkKernel(int streamIndex)
+void CudaStreamManager::triggerStartBulkKernel(CudaStreamIndex streamIndex, uint multiStreamIndex)
{
- checkCudaErrors(cudaEventRecord(startBulkKernel, cudaStreams[streamIndex]));
+ checkCudaErrors(cudaEventRecord(startBulkKernel, getStream(streamIndex, multiStreamIndex)));
}
-void CudaStreamManager::waitOnStartBulkKernelEvent(int streamIndex)
+void CudaStreamManager::waitOnStartBulkKernelEvent(CudaStreamIndex streamIndex, uint multiStreamIndex)
{
- checkCudaErrors(cudaStreamWaitEvent(cudaStreams[streamIndex], startBulkKernel));
+ checkCudaErrors(cudaStreamWaitEvent(getStream(streamIndex, multiStreamIndex), startBulkKernel));
}
diff --git a/src/gpu/VirtualFluids_GPU/Parameter/CudaStreamManager.h b/src/gpu/VirtualFluids_GPU/Parameter/CudaStreamManager.h
index c2d515ab5fe9c24388632a7ca9e1e4c78b7f1467..5c59bcd3a5e6178d6e70a63f803caf8e29f32604 100644
--- a/src/gpu/VirtualFluids_GPU/Parameter/CudaStreamManager.h
+++ b/src/gpu/VirtualFluids_GPU/Parameter/CudaStreamManager.h
@@ -30,32 +30,40 @@
#ifndef STREAM_MANAGER_H
#define STREAM_MANAGER_H
-#include <vector>
-#include "Core/DataTypes.h"
-
+#include <map>
+#include <cuda.h>
#include <cuda_runtime.h>
+#include "Core/DataTypes.h"
+enum class CudaStreamIndex
+ {
+ Legacy,
+ Bulk,
+ SubDomainBorder,
+ Precursor,
+ ActuatorFarm
+ };
class CudaStreamManager
-{
+{
private:
- std::vector<cudaStream_t> cudaStreams;
+ std::multimap<CudaStreamIndex, cudaStream_t> cudaStreams;
cudaEvent_t startBulkKernel = NULL;
- const int borderStreamIndex = 1;
- const int bulkStreamIndex = 0;
+ cudaStream_t legacyStream = CU_STREAM_LEGACY;
+
public:
- void launchStreams(uint numberOfStreams);
+ void registerStream(CudaStreamIndex streamIndex);
+ void launchStreams();
void terminateStreams();
- cudaStream_t &getStream(uint streamIndex);
-
- int getBorderStreamIndex();
- int getBulkStreamIndex();
+ cudaStream_t &getStream(CudaStreamIndex streamIndex, uint multiStreamIndex=0);
+ bool streamIsRegistered(CudaStreamIndex streamIndex);
// Events
void createCudaEvents();
void destroyCudaEvents();
- void triggerStartBulkKernel(int streamIndex);
- void waitOnStartBulkKernelEvent(int strteamIndex);
+
+ void triggerStartBulkKernel(CudaStreamIndex streamIndex, uint multiStreamIndex=0);
+ void waitOnStartBulkKernelEvent(CudaStreamIndex streamIndex, uint multiStreamIndex=0);
};
#endif
diff --git a/src/gpu/VirtualFluids_GPU/Parameter/Parameter.cpp b/src/gpu/VirtualFluids_GPU/Parameter/Parameter.cpp
index 7687ec926270f23e57608ca5f3084bd26d4de20e..d1568932d9802d1060e37d26ed37fa48de1abb05 100644
--- a/src/gpu/VirtualFluids_GPU/Parameter/Parameter.cpp
+++ b/src/gpu/VirtualFluids_GPU/Parameter/Parameter.cpp
@@ -65,6 +65,8 @@ Parameter::Parameter(int numberOfProcesses, int myId, std::optional<const vf::ba
initGridPaths();
initGridBasePoints();
initDefaultLBMkernelAllLevels();
+
+ this->cudaStreamManager = std::make_unique<CudaStreamManager>();
}
Parameter::~Parameter() = default;
@@ -829,7 +831,7 @@ real Parameter::getLengthRatio()
}
real Parameter::getForceRatio()
{
- return this->getDensityRatio() * this->getVelocityRatio()/this->getTimeRatio();
+ return (this->getDensityRatio()+1.0) * this->getVelocityRatio()/this->getTimeRatio();
}
real Parameter::getScaledViscosityRatio(int level)
{
@@ -859,6 +861,10 @@ real Parameter::getScaledForceRatio(int level)
{
return this->getForceRatio()*(level+1);
}
+real Parameter::getScaledStressRatio(int level)
+{
+ return this->getVelocityRatio()*this->getVelocityRatio();
+}
void Parameter::setRealX(real RealX)
{
ic.RealX = RealX;
@@ -883,6 +889,10 @@ void Parameter::setPressOutZ(unsigned int PressOutZ)
{
ic.PressOutZ = PressOutZ;
}
+void Parameter::setOutflowPressureCorrectionFactor(real pressBCrhoCorrectionFactor)
+{
+ ic.outflowPressureCorrectionFactor = pressBCrhoCorrectionFactor;
+}
void Parameter::setMaxDev(int maxdev)
{
ic.maxdev = maxdev;
@@ -1607,7 +1617,7 @@ void Parameter::setOutflowBoundaryNormalZ(std::string outflowNormalZ)
void Parameter::setMainKernel(std::string kernel)
{
this->mainKernel = kernel;
- if (kernel.find("Stream") != std::string::npos || kernel.find("Redesigned") != std::string::npos)
+ if ( kernel.find("CumulantK17") != std::string::npos )
this->kernelNeedsFluidNodeIndicesToRun = true;
}
void Parameter::setMultiKernelOn(bool isOn)
@@ -1916,6 +1926,10 @@ unsigned int Parameter::getPressOutZ()
{
return ic.PressOutZ;
}
+real Parameter::getOutflowPressureCorrectionFactor()
+{
+ return ic.outflowPressureCorrectionFactor;
+}
int Parameter::getMaxDev()
{
return ic.maxdev;
@@ -2657,8 +2671,7 @@ void Parameter::setUseStreams(bool useStreams)
if (useStreams) {
if (this->getNumprocs() != 1) {
this->useStreams = useStreams;
- this->cudaStreamManager = std::make_unique<CudaStreamManager>();
- return;
+ return;
} else {
std::cout << "Can't use streams with only one process!" << std::endl;
}
diff --git a/src/gpu/VirtualFluids_GPU/Parameter/Parameter.h b/src/gpu/VirtualFluids_GPU/Parameter/Parameter.h
index cbb8bfd68702bc2285947eb76e6d0adc54a5b6c1..86b7bc2a058d69ba878d4445953a4dc56e524027 100644
--- a/src/gpu/VirtualFluids_GPU/Parameter/Parameter.h
+++ b/src/gpu/VirtualFluids_GPU/Parameter/Parameter.h
@@ -54,6 +54,8 @@ class ConfigurationFile;
}
class CudaStreamManager;
+class TransientBCInputFileReader;
+
//! \struct LBMSimulationParameter
//! \brief struct holds and manages the LB-parameter of the simulation
//! \brief For this purpose it holds structures and pointer for host and device data, respectively.
@@ -218,16 +220,16 @@ struct LBMSimulationParameter {
OffsetFC offFCBulk;
unsigned int mem_size_kCF_off;
unsigned int mem_size_kFC_off;
-
- // BC's////////////////////
+
//! \brief stores the boundary condition data
QforBoundaryConditions noSlipBC, velocityBC, outflowBC, slipBC, stressBC, pressureBC;
//! \brief number of lattice nodes for the boundary conditions
- unsigned int numberOfNoSlipBCnodesRead, numberOfVeloBCnodesRead, numberOfOutflowBCnodesRead, numberOfSlipBCnodesRead, numberOfStressBCnodesRead, numberOfPressureBCnodesRead;
+ unsigned int numberOfNoSlipBCnodesRead, numberOfVeloBCnodesRead, numberOfOutflowBCnodesRead, numberOfSlipBCnodesRead, numberOfStressBCnodesRead, numberOfPressureBCnodesRead, numberOfPrecursorBCnodesRead;
QforBoundaryConditions QpressX0, QpressX1, QpressY0, QpressY1, QpressZ0, QpressZ1; // DEPRECATED
QforBoundaryConditions propellerBC;
QforBoundaryConditions geometryBC;
+ QforPrecursorBoundaryConditions precursorBC;
QforBoundaryConditions geometryBCnormalX, geometryBCnormalY, geometryBCnormalZ;
QforBoundaryConditions inflowBCnormalX, inflowBCnormalY, inflowBCnormalZ;
QforBoundaryConditions outflowBCnormalX, outflowBCnormalY, outflowBCnormalZ;
@@ -235,6 +237,8 @@ struct LBMSimulationParameter {
unsigned int kInletQread, kOutletQread; // DEPRECATED
WallModelParameters wallModel;
+ std::vector<SPtr<TransientBCInputFileReader>> transientBCInputFileReader;
+ real outflowPressureCorrectionFactor;
// testRoundoffError
Distributions27 kDistTestRE;
@@ -367,10 +371,19 @@ struct LBMSimulationParameter {
std::vector<EdgeNodePositions> edgeNodesYtoZ;
///////////////////////////////////////////////////////
- uint *fluidNodeIndices;
- uint numberOfFluidNodes;
- uint *fluidNodeIndicesBorder;
- uint numberOfFluidNodesBorder;
+ std::map<CollisionTemplate, uint*> taggedFluidNodeIndices = {{CollisionTemplate::Default, nullptr},
+ {CollisionTemplate::SubDomainBorder,nullptr},
+ {CollisionTemplate::WriteMacroVars, nullptr},
+ {CollisionTemplate::ApplyBodyForce, nullptr},
+ {CollisionTemplate::AllFeatures, nullptr}};
+ std::map<CollisionTemplate, uint > numberOfTaggedFluidNodes = {{CollisionTemplate::Default, 0},
+ {CollisionTemplate::SubDomainBorder,0},
+ {CollisionTemplate::WriteMacroVars, 0},
+ {CollisionTemplate::ApplyBodyForce, 0},
+ {CollisionTemplate::AllFeatures, 0}};
+
+ std::vector<CollisionTemplate> allocatedBulkFluidNodeTags = {};
+
};
//! \brief Class for LBM-parameter management
@@ -471,6 +484,7 @@ public:
void setpressBcPos(std::string pressBcPos);
void setpressBcQs(std::string pressBcQs);
void setpressBcValue(std::string pressBcValue);
+ void setOutflowPressureCorrectionFactor(real correctionFactor);
void setpressBcValues(std::string pressBcValues);
void setvelBcQs(std::string velBcQs);
void setvelBcValues(std::string velBcValues);
@@ -527,7 +541,6 @@ public:
void setUseWale(bool useWale);
void setTurbulenceModel(TurbulenceModel turbulenceModel);
void setUseTurbulentViscosity(bool useTurbulentViscosity);
- void setUseAMD(bool useAMD);
void setSGSConstant(real SGSConstant);
void setHasWallModelMonitor(bool hasWallModelMonitor);
void setUseInitNeq(bool useInitNeq);
@@ -765,6 +778,8 @@ public:
real getScaledDensityRatio(int level);
//! \returns the pressure ratio in SI/LB units scaled to the respective level
real getScaledPressureRatio(int level);
+ //! \returns the stress ratio in SI/LB units scaled to the respective level
+ real getScaledStressRatio(int level);
//! \returns the time ratio in SI/LB units scaled to the respective level
real getScaledTimeRatio(int level);
//! \returns the length ratio in SI/LB units scaled to the respective level
@@ -853,6 +868,7 @@ public:
std::string getOutflowBoundaryNormalX();
std::string getOutflowBoundaryNormalY();
std::string getOutflowBoundaryNormalZ();
+ real getOutflowPressureCorrectionFactor();
// CUDA random number
curandState *getRandomState();
// Kernel
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorFarm.cu b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorFarm.cu
new file mode 100644
index 0000000000000000000000000000000000000000..626e3bc0b745a8f8b029afb38028c81bd8d659e7
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorFarm.cu
@@ -0,0 +1,594 @@
+#include "ActuatorFarm.h"
+
+#include <cuda.h>
+#include <cuda_runtime.h>
+#include <helper_cuda.h>
+
+#include <cuda/CudaGrid.h>
+#include "VirtualFluids_GPU/GPU/GeometryUtils.h"
+#include "VirtualFluids_GPU/Kernel/Utilities/DistributionHelper.cuh"
+
+#include "Parameter/Parameter.h"
+#include "Parameter/CudaStreamManager.h"
+#include "DataStructureInitializer/GridProvider.h"
+#include "GPU/CudaMemoryManager.h"
+#include <lbm/constants/NumericConstants.h>
+#include <logger/Logger.h>
+
+using namespace vf::lbm::constant;
+
+
+__host__ __device__ __inline__ uint calcNode(uint bladeNode, uint numberOfBladeNodes, uint blade, uint numberOfBlades, uint turbine, uint numberOfTurbines)
+{
+
+ return bladeNode+numberOfBladeNodes*(blade+numberOfBlades*turbine);
+}
+
+__host__ __device__ __inline__ void calcTurbineBladeAndBladeNode(uint node, uint& bladeNode, uint numberOfBladeNodes, uint& blade, uint numberOfBlades, uint& turbine, uint numberOfTurbines)
+{
+ turbine = node/(numberOfBladeNodes*numberOfBlades);
+ uint x_off = turbine*numberOfBladeNodes*numberOfBlades;
+ blade = (node - x_off)/numberOfBlades;
+ uint y_off = numberOfBladeNodes*blade+x_off;
+ bladeNode = (node - y_off)/numberOfBladeNodes;
+}
+
+__host__ __device__ __forceinline__ real distSqrd(real distX, real distY, real distZ)
+{
+ return distX*distX+distY*distY+distZ*distZ;
+}
+
+void swapArrays(real* &arr1, real* &arr2)
+{
+ real* tmp = arr1;
+ arr1 = arr2;
+ arr2 = tmp;
+}
+
+__host__ __device__ __inline__ void rotateFromBladeToGlobal(
+ real& bladeCoordX_BF, real& bladeCoordY_BF, real& bladeCoordZ_BF,
+ real& bladeCoordX_GF, real& bladeCoordY_GF, real& bladeCoordZ_GF,
+ real& azimuth, real& yaw)
+{
+ real tmpX, tmpY, tmpZ;
+
+ rotateAboutX3D(azimuth, bladeCoordX_BF, bladeCoordY_BF, bladeCoordZ_BF, tmpX, tmpY, tmpZ);
+ rotateAboutZ3D(yaw, tmpX, tmpY, tmpZ, bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF);
+
+}
+
+__host__ __device__ __inline__ void rotateFromGlobalToBlade(
+ real& bladeCoordX_BF, real& bladeCoordY_BF, real& bladeCoordZ_BF,
+ real& bladeCoordX_GF, real& bladeCoordY_GF, real& bladeCoordZ_GF,
+ real& azimuth, real& yaw)
+{
+ real tmpX, tmpY, tmpZ;
+
+ invRotateAboutZ3D(yaw, bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF, tmpX, tmpY, tmpZ);
+ invRotateAboutX3D(azimuth, tmpX, tmpY, tmpZ, bladeCoordX_BF, bladeCoordY_BF, bladeCoordZ_BF);
+}
+
+__global__ void interpolateVelocities(real* gridCoordsX, real* gridCoordsY, real* gridCoordsZ,
+ uint* neighborsX, uint* neighborsY, uint* neighborsZ, uint* neighborsWSB,
+ real* vx, real* vy, real* vz,
+ real* bladeCoordsX, real* bladeCoordsY, real* bladeCoordsZ,
+ real* bladeVelocitiesX, real* bladeVelocitiesY, real* bladeVelocitiesZ,
+ uint numberOfTurbines, uint numberOfBlades, uint numberOfBladeNodes,
+ real* azimuths, real* yaws, real* omegas,
+ real* turbPosX, real* turbPosY, real* turbPosZ,
+ uint* bladeIndices, real velocityRatio, real invDeltaX)
+{
+
+ const uint node = vf::gpu::getNodeIndex();
+
+ if(node>=numberOfBladeNodes*numberOfBlades*numberOfTurbines) return;
+
+ uint turbine, bladeNode, blade;
+
+ calcTurbineBladeAndBladeNode(node, bladeNode, numberOfBladeNodes, blade, numberOfBlades, turbine, numberOfTurbines);
+
+ real bladeCoordX_BF = bladeCoordsX[node];
+ real bladeCoordY_BF = bladeCoordsY[node];
+ real bladeCoordZ_BF = bladeCoordsZ[node];
+
+ real bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF;
+
+ real localAzimuth = azimuths[turbine]+blade*c2Pi/numberOfBlades;
+ real yaw = yaws[turbine];
+
+
+ rotateFromBladeToGlobal(bladeCoordX_BF, bladeCoordY_BF, bladeCoordZ_BF,
+ bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF,
+ localAzimuth, yaw);
+
+ bladeCoordX_GF += turbPosX[turbine];
+ bladeCoordY_GF += turbPosY[turbine];
+ bladeCoordZ_GF += turbPosZ[turbine];
+
+ uint k, ke, kn, kt;
+ uint kne, kte, ktn, ktne;
+
+ k = findNearestCellBSW(bladeIndices[node],
+ gridCoordsX, gridCoordsY, gridCoordsZ,
+ bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF,
+ neighborsX, neighborsY, neighborsZ, neighborsWSB);
+
+ bladeIndices[node] = k;
+
+ getNeighborIndicesOfBSW(k, ke, kn, kt, kne, kte, ktn, ktne, neighborsX, neighborsY, neighborsZ);
+
+ real dW, dE, dN, dS, dT, dB;
+
+ real distX = invDeltaX*(bladeCoordX_GF-gridCoordsX[k]);
+ real distY = invDeltaX*(bladeCoordY_GF-gridCoordsY[k]);
+ real distZ = invDeltaX*(bladeCoordZ_GF-gridCoordsZ[k]);
+
+ getInterpolationWeights(dW, dE, dN, dS, dT, dB, distX, distY, distZ);
+
+ real bladeVelX_GF = trilinearInterpolation(dW, dE, dN, dS, dT, dB, k, ke, kn, kt, kne, kte, ktn, ktne, vx)*velocityRatio;
+ real bladeVelY_GF = trilinearInterpolation(dW, dE, dN, dS, dT, dB, k, ke, kn, kt, kne, kte, ktn, ktne, vy)*velocityRatio;
+ real bladeVelZ_GF = trilinearInterpolation(dW, dE, dN, dS, dT, dB, k, ke, kn, kt, kne, kte, ktn, ktne, vz)*velocityRatio;
+
+ real bladeVelX_BF, bladeVelY_BF, bladeVelZ_BF;
+
+ rotateFromGlobalToBlade(bladeVelX_BF, bladeVelY_BF, bladeVelZ_BF,
+ bladeVelX_GF, bladeVelY_GF, bladeVelZ_GF,
+ localAzimuth, yaw);
+
+ bladeVelocitiesX[node] = bladeVelX_BF;
+ bladeVelocitiesY[node] = bladeVelY_BF+omegas[turbine]*bladeCoordZ_BF;
+ bladeVelocitiesZ[node] = bladeVelZ_BF;
+}
+
+
+__global__ void applyBodyForces(real* gridCoordsX, real* gridCoordsY, real* gridCoordsZ,
+ real* gridForcesX, real* gridForcesY, real* gridForcesZ,
+ real* bladeCoordsX, real* bladeCoordsY, real* bladeCoordsZ,
+ real* bladeForcesX, real* bladeForcesY, real* bladeForcesZ,
+ const uint numberOfTurbines, const uint numberOfBlades, const uint numberOfBladeNodes,
+ real* azimuths, real* yaws, real* diameters,
+ real* turbPosX, real* turbPosY, real* turbPosZ,
+ uint* gridIndices, uint nIndices,
+ const real invEpsilonSqrd, const real factorGaussian)
+{
+
+ const uint index = vf::gpu::getNodeIndex();
+
+ if(index>=nIndices) return;
+
+
+ uint gridIndex = gridIndices[index];
+
+ real gridCoordX_GF = gridCoordsX[gridIndex];
+ real gridCoordY_GF = gridCoordsY[gridIndex];
+ real gridCoordZ_GF = gridCoordsZ[gridIndex];
+
+ real gridForceX_RF = c0o1;
+ real gridForceY_RF = c0o1;
+ real gridForceZ_RF = c0o1;
+
+ real dAzimuth = c2Pi/numberOfBlades;
+
+ for(uint turbine = 0; turbine<numberOfTurbines; turbine++)
+ {
+ real radius = c1o2*diameters[turbine];
+ real gridCoordX_RF = gridCoordX_GF - turbPosX[turbine];
+ real gridCoordY_RF = gridCoordY_GF - turbPosY[turbine];
+ real gridCoordZ_RF = gridCoordZ_GF - turbPosZ[turbine];
+
+ if(distSqrd(gridCoordX_RF, gridCoordY_RF, gridCoordZ_RF)*invEpsilonSqrd > radius*radius*invEpsilonSqrd+c7o1)
+ continue;
+
+ real azimuth = azimuths[turbine];
+ real yaw = yaws[turbine];
+
+ for( uint blade=0; blade<numberOfBlades; blade++)
+ {
+ real localAzimuth = azimuth+blade*dAzimuth;
+
+
+ real gridCoordX_BF, gridCoordY_BF, gridCoordZ_BF;
+
+ rotateFromGlobalToBlade(gridCoordX_BF, gridCoordY_BF, gridCoordZ_BF,
+ gridCoordX_RF, gridCoordY_RF, gridCoordZ_RF,
+ localAzimuth, yaw);
+
+ uint node;
+ uint nextNode = calcNode(0, numberOfBladeNodes, blade, numberOfBlades, turbine, numberOfTurbines);
+
+ real last_z = c0o1;
+ real current_z = c0o1;
+ real next_z = bladeCoordsZ[nextNode];
+
+ real x, y, dz, eta, forceX_RF, forceY_RF, forceZ_RF;
+
+ for( uint bladeNode=0; bladeNode<numberOfBladeNodes-1; bladeNode++)
+ {
+ node = nextNode;
+ nextNode = calcNode(bladeNode+1, numberOfBladeNodes, blade, numberOfBlades, turbine, numberOfTurbines);
+
+ x = bladeCoordsX[node];
+ y = bladeCoordsY[node];
+ last_z = current_z;
+ current_z = next_z;
+ next_z = bladeCoordsZ[nextNode];
+
+ dz = c1o2*(next_z-last_z);
+
+ eta = dz*factorGaussian*exp(-distSqrd(x-gridCoordX_BF, y-gridCoordY_BF, current_z-gridCoordZ_BF)*invEpsilonSqrd);
+ rotateFromBladeToGlobal(bladeForcesX[node], bladeForcesY[node], bladeForcesZ[node],
+ forceX_RF, forceY_RF, forceZ_RF,
+ localAzimuth, yaw);
+
+ gridForceX_RF += forceX_RF*eta;
+ gridForceY_RF += forceY_RF*eta;
+ gridForceZ_RF += forceZ_RF*eta;
+ }
+
+ //Handle last node separately
+
+ node = nextNode;
+
+ x = bladeCoordsX[node];
+ y = bladeCoordsY[node];
+ last_z = current_z;
+ current_z = next_z;
+
+ dz = c1o2*(radius-last_z);
+
+ eta = dz*factorGaussian*exp(-distSqrd(x-gridCoordX_BF, y-gridCoordY_BF, current_z-gridCoordZ_BF)*invEpsilonSqrd);
+
+ rotateFromBladeToGlobal(bladeForcesX[node], bladeForcesY[node], bladeForcesZ[node],
+ forceX_RF, forceY_RF, forceZ_RF,
+ localAzimuth, yaw);
+
+ gridForceX_RF += forceX_RF*eta;
+ gridForceY_RF += forceY_RF*eta;
+ gridForceZ_RF += forceZ_RF*eta;
+ }
+ }
+
+ gridForcesX[gridIndex] += gridForceX_RF;
+ gridForcesY[gridIndex] += gridForceY_RF;
+ gridForcesZ[gridIndex] += gridForceZ_RF;
+}
+
+void ActuatorFarm::addTurbine(real posX, real posY, real posZ, real diameter, real omega, real azimuth, real yaw, std::vector<real> bladeRadii)
+{
+ preInitPosX.push_back(posX);
+ preInitPosY.push_back(posY);
+ preInitPosZ.push_back(posZ);
+ preInitOmegas.push_back(omega);
+ preInitAzimuths.push_back(azimuth);
+ preInitYaws.push_back(yaw);
+ preInitDiameters.push_back(diameter);
+ preInitBladeRadii.push_back(bladeRadii);
+}
+
+void ActuatorFarm::init(Parameter* para, GridProvider* gridProvider, CudaMemoryManager* cudaMemoryManager)
+{
+ if(!para->getIsBodyForce()) throw std::runtime_error("try to allocate ActuatorFarm but BodyForce is not set in Parameter.");
+ this->forceRatio = para->getForceRatio();
+ this->initTurbineGeometries(cudaMemoryManager);
+ this->initBladeCoords(cudaMemoryManager);
+ this->initBladeIndices(para, cudaMemoryManager);
+ this->initBladeVelocities(cudaMemoryManager);
+ this->initBladeForces(cudaMemoryManager);
+ this->initBoundingSpheres(para, cudaMemoryManager);
+ this->streamIndex = 0;
+}
+
+void ActuatorFarm::interact(Parameter* para, CudaMemoryManager* cudaMemoryManager, int level, unsigned int t)
+{
+ if (level != this->level) return;
+
+ cudaStream_t stream = para->getStreamManager()->getStream(CudaStreamIndex::ActuatorFarm, this->streamIndex);
+
+ if(useHostArrays) cudaMemoryManager->cudaCopyBladeCoordsHtoD(this);
+
+ vf::cuda::CudaGrid bladeGrid = vf::cuda::CudaGrid(para->getParH(level)->numberofthreads, this->numberOfNodes);
+
+ interpolateVelocities<<< bladeGrid.grid, bladeGrid.threads, 0, stream >>>(
+ para->getParD(this->level)->coordinateX, para->getParD(this->level)->coordinateY, para->getParD(this->level)->coordinateZ,
+ para->getParD(this->level)->neighborX, para->getParD(this->level)->neighborY, para->getParD(this->level)->neighborZ, para->getParD(this->level)->neighborInverse,
+ para->getParD(this->level)->velocityX, para->getParD(this->level)->velocityY, para->getParD(this->level)->velocityZ,
+ this->bladeCoordsXDCurrentTimestep, this->bladeCoordsYDCurrentTimestep, this->bladeCoordsZDCurrentTimestep,
+ this->bladeVelocitiesXDCurrentTimestep, this->bladeVelocitiesYDCurrentTimestep, this->bladeVelocitiesZDCurrentTimestep,
+ this->numberOfTurbines, this->numberOfBlades, this->numberOfBladeNodes,
+ this->azimuthsD, this->yawsD, this->omegasD,
+ this->turbinePosXD, this->turbinePosYD, this->turbinePosZD,
+ this->bladeIndicesD, para->getVelocityRatio(), this->invDeltaX);
+
+ cudaStreamSynchronize(stream);
+ if(useHostArrays) cudaMemoryManager->cudaCopyBladeVelocitiesDtoH(this);
+ this->calcBladeForces();
+ this->swapDeviceArrays();
+
+ if(useHostArrays) cudaMemoryManager->cudaCopyBladeForcesHtoD(this);
+
+ vf::cuda::CudaGrid sphereGrid = vf::cuda::CudaGrid(para->getParH(level)->numberofthreads, this->numberOfIndices);
+
+ applyBodyForces<<<sphereGrid.grid, sphereGrid.threads, 0, stream>>>(
+ para->getParD(this->level)->coordinateX, para->getParD(this->level)->coordinateY, para->getParD(this->level)->coordinateZ,
+ para->getParD(this->level)->forceX_SP, para->getParD(this->level)->forceY_SP, para->getParD(this->level)->forceZ_SP,
+ this->bladeCoordsXDCurrentTimestep, this->bladeCoordsYDCurrentTimestep, this->bladeCoordsZDCurrentTimestep,
+ this->bladeForcesXDCurrentTimestep, this->bladeForcesYDCurrentTimestep, this->bladeForcesZDCurrentTimestep,
+ this->numberOfTurbines, this->numberOfBlades, this->numberOfBladeNodes,
+ this->azimuthsD, this->yawsD, this->diametersD,
+ this->turbinePosXD, this->turbinePosYD, this->turbinePosZD,
+ this->boundingSphereIndicesD, this->numberOfIndices,
+ this->invEpsilonSqrd, this->factorGaussian);
+ cudaMemoryManager->cudaCopyBladeOrientationsHtoD(this);
+ cudaStreamSynchronize(stream);
+}
+
+
+void ActuatorFarm::free(Parameter* para, CudaMemoryManager* cudaMemoryManager)
+{
+ cudaMemoryManager->cudaFreeBladeGeometries(this);
+ cudaMemoryManager->cudaFreeBladeOrientations(this);
+ cudaMemoryManager->cudaFreeBladeCoords(this);
+ cudaMemoryManager->cudaFreeBladeVelocities(this);
+ cudaMemoryManager->cudaFreeBladeForces(this);
+ cudaMemoryManager->cudaFreeBladeIndices(this);
+ cudaMemoryManager->cudaFreeSphereIndices(this);
+}
+
+
+void ActuatorFarm::calcForcesEllipticWing()
+{
+ real u_rel, v_rel, u_rel_sq;
+ real phi;
+ real Cl = c1o1;
+ real Cd = c0o1;
+ real c0 = 20*c1o10;
+ real c, Cn, Ct;
+ for(uint turbine=0; turbine<this->numberOfTurbines; turbine++)
+ {
+ real diameter = this->diametersH[turbine];
+ for( uint blade=0; blade<this->numberOfBlades; blade++)
+ {
+ for( uint bladeNode=0; bladeNode<this->numberOfBladeNodes; bladeNode++)
+ {
+ uint node = calcNode(bladeNode, this->numberOfBladeNodes, blade, this->numberOfBlades, turbine, this->numberOfTurbines);
+
+ u_rel = this->bladeVelocitiesXH[node];
+ v_rel = this->bladeVelocitiesYH[node];
+ u_rel_sq = u_rel*u_rel+v_rel*v_rel;
+ phi = atan2(u_rel, v_rel);
+
+ real tmp = c4o1*this->bladeRadiiH[bladeNode]/diameter-c1o1;
+ c = c0 * sqrt( c1o1- tmp*tmp );
+ Cn = Cl*cos(phi)+Cd*sin(phi);
+ Ct = Cl*sin(phi)-Cd*cos(phi);
+ real fx = c1o2*u_rel_sq*c*this->density*Cn;
+ real fy = c1o2*u_rel_sq*c*this->density*Ct;
+ this->bladeForcesXH[node] = -fx;
+ this->bladeForcesYH[node] = -fy;
+ this->bladeForcesZH[node] = c0o1;
+ // printf("u %f v %f fx %f fy %f \n", u_rel, v_rel, fx, fy);
+ }
+ }
+ azimuthsH[turbine] = azimuthsH[turbine]+deltaT*omegasH[turbine];
+ }
+}
+
+void ActuatorFarm::calcBladeForces()
+{
+ this->calcForcesEllipticWing();
+}
+
+void ActuatorFarm::getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider)
+{
+ std::vector<uint> indicesInSphere(this->boundingSphereIndicesH, this->boundingSphereIndicesH+this->numberOfIndices);
+ gridProvider->tagFluidNodeIndices(indicesInSphere, CollisionTemplate::AllFeatures, this->level);
+}
+
+
+void ActuatorFarm::initTurbineGeometries(CudaMemoryManager* cudaMemoryManager)
+{
+ this->numberOfTurbines = uint(this->preInitDiameters.size());
+ this->numberOfNodes = numberOfTurbines*numberOfBladeNodes*numberOfBlades;
+
+ cudaMemoryManager->cudaAllocBladeGeometries(this);
+ cudaMemoryManager->cudaAllocBladeOrientations(this);
+
+ for(uint turbine=0; turbine<this->numberOfTurbines; turbine++)
+ {
+ for(uint node=0; node<this->numberOfBladeNodes; node++)
+ {
+ this->bladeRadiiH[calcNode(node, numberOfBladeNodes, 0, 1, turbine, numberOfTurbines)] = this->preInitBladeRadii[turbine][node];
+ }
+
+ }
+ std::copy(preInitPosX.begin(), preInitPosX.end(), turbinePosXH);
+ std::copy(preInitPosY.begin(), preInitPosY.end(), turbinePosYH);
+ std::copy(preInitPosZ.begin(), preInitPosZ.end(), turbinePosZH);
+ std::copy(preInitDiameters.begin(), preInitDiameters.end(), diametersH);
+
+ cudaMemoryManager->cudaCopyBladeGeometriesHtoD(this);
+ std::copy(preInitAzimuths.begin(), preInitAzimuths.end(), this->azimuthsH);
+ std::copy(preInitOmegas.begin(), preInitOmegas.end(), this->omegasH);
+ std::copy(preInitYaws.begin(), preInitYaws.end(), this->yawsH);
+
+ cudaMemoryManager->cudaCopyBladeOrientationsHtoD(this);
+ this->factorGaussian = pow(this->epsilon*sqrt(cPi),-c3o1)/this->forceRatio;
+}
+
+void ActuatorFarm::initBladeCoords(CudaMemoryManager* cudaMemoryManager)
+{
+ cudaMemoryManager->cudaAllocBladeCoords(this);
+
+ for(uint turbine=0; turbine<numberOfTurbines; turbine++)
+ {
+ for(uint blade=0; blade<this->numberOfBlades; blade++)
+ {
+ for(uint bladeNode=0; bladeNode<this->numberOfBladeNodes; bladeNode++)
+ {
+ uint node = calcNode(bladeNode, this->numberOfBladeNodes, blade, this->numberOfBlades, turbine, this->numberOfTurbines);
+
+ this->bladeCoordsXH[node] = c0o1;
+ this->bladeCoordsYH[node] = c0o1;
+ this->bladeCoordsZH[node] = this->bladeRadiiH[calcNode(bladeNode, numberOfBladeNodes, 0, 1, turbine, numberOfTurbines)];
+ }
+ }
+ }
+ cudaMemoryManager->cudaCopyBladeCoordsHtoD(this);
+ swapArrays(this->bladeCoordsXDCurrentTimestep, this->bladeCoordsXDPreviousTimestep);
+ swapArrays(this->bladeCoordsYDCurrentTimestep, this->bladeCoordsYDPreviousTimestep);
+ swapArrays(this->bladeCoordsZDCurrentTimestep, this->bladeCoordsZDPreviousTimestep);
+ cudaMemoryManager->cudaCopyBladeCoordsHtoD(this);
+}
+
+void ActuatorFarm::initBladeVelocities(CudaMemoryManager* cudaMemoryManager)
+{
+ cudaMemoryManager->cudaAllocBladeVelocities(this);
+
+ std::fill_n(this->bladeVelocitiesXH, this->numberOfNodes, c0o1);
+ std::fill_n(this->bladeVelocitiesYH, this->numberOfNodes, c0o1);
+ std::fill_n(this->bladeVelocitiesZH, this->numberOfNodes, c0o1);
+
+ cudaMemoryManager->cudaCopyBladeVelocitiesHtoD(this);
+ swapArrays(this->bladeVelocitiesXDCurrentTimestep, this->bladeVelocitiesXDPreviousTimestep);
+ swapArrays(this->bladeVelocitiesYDCurrentTimestep, this->bladeVelocitiesYDPreviousTimestep);
+ swapArrays(this->bladeVelocitiesZDCurrentTimestep, this->bladeVelocitiesZDPreviousTimestep);
+ cudaMemoryManager->cudaCopyBladeVelocitiesHtoD(this);
+}
+
+void ActuatorFarm::initBladeForces(CudaMemoryManager* cudaMemoryManager)
+{
+ cudaMemoryManager->cudaAllocBladeForces(this);
+
+ std::fill_n(this->bladeForcesXH, this->numberOfNodes, c0o1);
+ std::fill_n(this->bladeForcesYH, this->numberOfNodes, c0o1);
+ std::fill_n(this->bladeForcesZH, this->numberOfNodes, c0o1);
+
+ cudaMemoryManager->cudaCopyBladeForcesHtoD(this);
+ swapArrays(this->bladeForcesXDCurrentTimestep, this->bladeForcesXDPreviousTimestep);
+ swapArrays(this->bladeForcesYDCurrentTimestep, this->bladeForcesYDPreviousTimestep);
+ swapArrays(this->bladeForcesZDCurrentTimestep, this->bladeForcesZDPreviousTimestep);
+ cudaMemoryManager->cudaCopyBladeForcesHtoD(this);
+}
+
+void ActuatorFarm::initBladeIndices(Parameter* para, CudaMemoryManager* cudaMemoryManager)
+{
+ cudaMemoryManager->cudaAllocBladeIndices(this);
+
+ std::fill_n(this->bladeIndicesH, this->numberOfNodes, 1);
+
+ cudaMemoryManager->cudaCopyBladeIndicesHtoD(this);
+}
+
+void ActuatorFarm::initBoundingSpheres(Parameter* para, CudaMemoryManager* cudaMemoryManager)
+{
+ std::vector<int> nodesInSpheres;
+
+ for(uint turbine=0; turbine<this->numberOfTurbines; turbine++)
+ {
+ real sphereRadius = c1o2*this->diametersH[turbine]+c4o1*this->epsilon;
+
+ real posX = this->turbinePosXH[turbine];
+ real posY = this->turbinePosYH[turbine];
+ real posZ = this->turbinePosZH[turbine];
+
+ real sphereRadiusSqrd = sphereRadius*sphereRadius;
+
+ uint minimumNumberOfNodesPerSphere = (uint)(c4o3*cPi*pow(sphereRadius-this->deltaX, c3o1)/pow(this->deltaX, c3o1));
+ uint nodesInThisSphere = 0;
+
+ for (uint j = 1; j <= para->getParH(this->level)->numberOfNodes; j++)
+ {
+ const real distX = para->getParH(this->level)->coordinateX[j]-posX;
+ const real distY = para->getParH(this->level)->coordinateY[j]-posY;
+ const real distZ = para->getParH(this->level)->coordinateZ[j]-posZ;
+ if(distSqrd(distX,distY,distZ) < sphereRadiusSqrd)
+ {
+ nodesInSpheres.push_back(j);
+ nodesInThisSphere++;
+ }
+ }
+
+ if(nodesInThisSphere<minimumNumberOfNodesPerSphere)
+ {
+ VF_LOG_CRITICAL("Found only {} nodes in bounding sphere of turbine no. {}, expected at least {}!", nodesInThisSphere, turbine, minimumNumberOfNodesPerSphere);
+ throw std::runtime_error("ActuatorFarm::initBoundingSpheres: Turbine bounding sphere partially out of domain.");
+ }
+ }
+
+ this->numberOfIndices = uint(nodesInSpheres.size());
+
+ cudaMemoryManager->cudaAllocSphereIndices(this);
+ std::copy(nodesInSpheres.begin(), nodesInSpheres.end(), this->boundingSphereIndicesH);
+ cudaMemoryManager->cudaCopySphereIndicesHtoD(this);
+}
+
+void ActuatorFarm::setAllAzimuths(real* _azimuths)
+{
+ std::copy_n(_azimuths, this->numberOfTurbines, this->azimuthsH);
+}
+
+void ActuatorFarm::setAllOmegas(real* _omegas)
+{
+ std::copy_n(_omegas, this->numberOfTurbines, this->omegasH);
+}
+
+void ActuatorFarm::setAllYaws(real* _yaws)
+{
+ std::copy_n(_yaws, this->numberOfTurbines, this->yawsH);
+}
+
+void ActuatorFarm::setAllBladeCoords(real* _bladeCoordsX, real* _bladeCoordsY, real* _bladeCoordsZ)
+{
+ std::copy_n(_bladeCoordsX, this->numberOfNodes, this->bladeCoordsXH);
+ std::copy_n(_bladeCoordsY, this->numberOfNodes, this->bladeCoordsYH);
+ std::copy_n(_bladeCoordsZ, this->numberOfNodes, this->bladeCoordsZH);
+}
+
+void ActuatorFarm::setAllBladeVelocities(real* _bladeVelocitiesX, real* _bladeVelocitiesY, real* _bladeVelocitiesZ)
+{
+ std::copy_n(_bladeVelocitiesX, this->numberOfNodes, this->bladeVelocitiesXH);
+ std::copy_n(_bladeVelocitiesY, this->numberOfNodes, this->bladeVelocitiesYH);
+ std::copy_n(_bladeVelocitiesZ, this->numberOfNodes, this->bladeVelocitiesZH);
+}
+
+void ActuatorFarm::setAllBladeForces(real* _bladeForcesX, real* _bladeForcesY, real* _bladeForcesZ)
+{
+ std::copy_n(_bladeForcesX, this->numberOfNodes, this->bladeForcesXH);
+ std::copy_n(_bladeForcesY, this->numberOfNodes, this->bladeForcesYH);
+ std::copy_n(_bladeForcesZ, this->numberOfNodes, this->bladeForcesZH);
+
+}void ActuatorFarm::setTurbineBladeCoords(uint turbine, real* _bladeCoordsX, real* _bladeCoordsY, real* _bladeCoordsZ)
+{
+ std::copy_n(_bladeCoordsX, numberOfBladeNodes*numberOfBlades, &this->bladeCoordsXH[turbine*numberOfBladeNodes*numberOfBlades]);
+ std::copy_n(_bladeCoordsY, numberOfBladeNodes*numberOfBlades, &this->bladeCoordsYH[turbine*numberOfBladeNodes*numberOfBlades]);
+ std::copy_n(_bladeCoordsZ, numberOfBladeNodes*numberOfBlades, &this->bladeCoordsZH[turbine*numberOfBladeNodes*numberOfBlades]);
+}
+
+void ActuatorFarm::setTurbineBladeVelocities(uint turbine, real* _bladeVelocitiesX, real* _bladeVelocitiesY, real* _bladeVelocitiesZ)
+{
+ std::copy_n(_bladeVelocitiesX, numberOfBladeNodes*numberOfBlades, &this->bladeVelocitiesXH[turbine*numberOfBladeNodes*numberOfBlades]);
+ std::copy_n(_bladeVelocitiesY, numberOfBladeNodes*numberOfBlades, &this->bladeVelocitiesYH[turbine*numberOfBladeNodes*numberOfBlades]);
+ std::copy_n(_bladeVelocitiesZ, numberOfBladeNodes*numberOfBlades, &this->bladeVelocitiesZH[turbine*numberOfBladeNodes*numberOfBlades]);
+}
+
+void ActuatorFarm::setTurbineBladeForces(uint turbine, real* _bladeForcesX, real* _bladeForcesY, real* _bladeForcesZ)
+{
+ std::copy_n(_bladeForcesX, numberOfBladeNodes*numberOfBlades, &this->bladeForcesXH[turbine*numberOfBladeNodes*numberOfBlades]);
+ std::copy_n(_bladeForcesY, numberOfBladeNodes*numberOfBlades, &this->bladeForcesYH[turbine*numberOfBladeNodes*numberOfBlades]);
+ std::copy_n(_bladeForcesZ, numberOfBladeNodes*numberOfBlades, &this->bladeForcesZH[turbine*numberOfBladeNodes*numberOfBlades]);
+}
+
+void ActuatorFarm::swapDeviceArrays()
+{
+ swapArrays(this->bladeCoordsXDPreviousTimestep, this->bladeCoordsXDCurrentTimestep);
+ swapArrays(this->bladeCoordsYDPreviousTimestep, this->bladeCoordsYDCurrentTimestep);
+ swapArrays(this->bladeCoordsZDPreviousTimestep, this->bladeCoordsZDCurrentTimestep);
+
+ swapArrays(this->bladeVelocitiesXDPreviousTimestep, this->bladeVelocitiesXDCurrentTimestep);
+ swapArrays(this->bladeVelocitiesYDPreviousTimestep, this->bladeVelocitiesYDCurrentTimestep);
+ swapArrays(this->bladeVelocitiesZDPreviousTimestep, this->bladeVelocitiesZDCurrentTimestep);
+
+ swapArrays(this->bladeForcesXDPreviousTimestep, this->bladeForcesXDCurrentTimestep);
+ swapArrays(this->bladeForcesYDPreviousTimestep, this->bladeForcesYDCurrentTimestep);
+ swapArrays(this->bladeForcesZDPreviousTimestep, this->bladeForcesZDCurrentTimestep);
+}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorFarm.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorFarm.h
new file mode 100644
index 0000000000000000000000000000000000000000..f771ce10cdfe81237ae1caa8a26f887488839aa4
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorFarm.h
@@ -0,0 +1,200 @@
+#ifndef ActuatorFarm_H
+#define ActuatorFarm_H
+
+#include "PreCollisionInteractor.h"
+#include "PointerDefinitions.h"
+#include "lbm/constants/NumericConstants.h"
+#include <stdexcept>
+
+using namespace vf::lbm::constant;
+
+class Parameter;
+class GridProvider;
+using namespace vf::lbm::constant;
+
+class ActuatorFarm : public PreCollisionInteractor
+{
+public:
+ ActuatorFarm(
+ const uint _nBlades,
+ const real _density,
+ const uint _nBladeNodes,
+ const real _epsilon,
+ int _level,
+ const real _deltaT,
+ const real _deltaX,
+ const bool _useHostArrays
+ ) :
+ numberOfBlades(_nBlades),
+ density(_density),
+ numberOfBladeNodes(_nBladeNodes),
+ epsilon(_epsilon),
+ level(_level),
+ useHostArrays(_useHostArrays),
+ numberOfTurbines(0),
+ numberOfNodes(0),
+ PreCollisionInteractor()
+ {
+ this->deltaT = _deltaT*exp2(-this->level);
+ this->deltaX = _deltaX*exp2(-this->level);
+ this->invEpsilonSqrd = 1/(epsilon*epsilon);
+ this->invDeltaX = c1o1/this->deltaX;
+
+ if(this->epsilon<this->deltaX)
+ throw std::runtime_error("ActuatorFarm::ActuatorFarm: epsilon needs to be larger than dx!");
+ }
+
+ virtual ~ActuatorFarm()
+ {
+
+ }
+ void addTurbine(real turbinePosX, real turbinePosY, real turbinePosZ, real diameter, real omega, real azimuth, real yaw, std::vector<real> bladeRadii);
+ void init(Parameter* para, GridProvider* gridProvider, CudaMemoryManager* cudaManager) override;
+ void interact(Parameter* para, CudaMemoryManager* cudaManager, int level, uint t) override;
+ void free(Parameter* para, CudaMemoryManager* cudaManager) override;
+ void getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider) override;
+
+ void write(uint t);
+
+ real getDensity(){ return this->density; };
+ real getDeltaT(){ return this->deltaT; };
+ real getDeltaX(){ return this->deltaX; };
+
+ uint getNumberOfTurbines(){ return this->numberOfTurbines; };
+ uint getNumberOfNodesPerBlade(){ return this->numberOfBladeNodes; };
+ uint getNumberOfBladesPerTurbine(){ return this->numberOfBlades; };
+
+ uint getNumberOfIndices(){ return this->numberOfIndices; };
+ uint getNumberOfNodes(){ return this->numberOfNodes; };
+
+ real* getAllAzimuths(){ return azimuthsH; };
+ real* getAllOmegas(){ return omegasH; };
+ real* getAllYaws(){ return yawsH; };
+
+ real* getAllTurbinePosX(){ return turbinePosXH; };
+ real* getAllTurbinePosY(){ return turbinePosYH; };
+ real* getAllTurbinePosZ(){ return turbinePosZH; };
+
+ real getTurbineAzimuth(uint turbine){ return azimuthsH[turbine]; };
+ real getTurbineOmega (uint turbine){ return omegasH[turbine]; };
+ real getTurbineYaw (uint turbine){ return yawsH[turbine]; };
+
+ real getTurbinePosX(uint turbine){ return turbinePosXH[turbine]; };
+ real getTurbinePosY(uint turbine){ return turbinePosYH[turbine]; };
+ real getTurbinePosZ(uint turbine){ return turbinePosZH[turbine]; };
+
+ real* getAllBladeRadii(){ return this->bladeRadiiH; };
+ real* getAllBladeCoordsX(){ return this->bladeCoordsXH; };
+ real* getAllBladeCoordsY(){ return this->bladeCoordsYH; };
+ real* getAllBladeCoordsZ(){ return this->bladeCoordsZH; };
+ real* getAllBladeVelocitiesX(){ return this->bladeVelocitiesXH; };
+ real* getAllBladeVelocitiesY(){ return this->bladeVelocitiesYH; };
+ real* getAllBladeVelocitiesZ(){ return this->bladeVelocitiesZH; };
+ real* getAllBladeForcesX(){ return this->bladeForcesXH; };
+ real* getAllBladeForcesY(){ return this->bladeForcesYH; };
+ real* getAllBladeForcesZ(){ return this->bladeForcesZH; };
+
+ real* getTurbineBladeRadii(uint turbine){ return &this->bladeRadiiH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeCoordsX(uint turbine){ return &this->bladeCoordsXH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeCoordsY(uint turbine){ return &this->bladeCoordsYH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeCoordsZ(uint turbine){ return &this->bladeCoordsZH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeVelocitiesX(uint turbine){ return &this->bladeVelocitiesXH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeVelocitiesY(uint turbine){ return &this->bladeVelocitiesYH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeVelocitiesZ(uint turbine){ return &this->bladeVelocitiesZH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeForcesX(uint turbine){ return &this->bladeForcesXH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeForcesY(uint turbine){ return &this->bladeForcesYH[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeForcesZ(uint turbine){ return &this->bladeForcesZH[turbine*numberOfBladeNodes*numberOfBlades]; };
+
+ real* getAllBladeRadiiDevice(){ return this->bladeRadiiD; };
+ real* getAllBladeCoordsXDevice(){ return this->bladeCoordsXDCurrentTimestep; };
+ real* getAllBladeCoordsYDevice(){ return this->bladeCoordsYDCurrentTimestep; };
+ real* getAllBladeCoordsZDevice(){ return this->bladeCoordsZDCurrentTimestep; };
+ real* getAllBladeVelocitiesXDevice(){ return this->bladeVelocitiesXDCurrentTimestep; };
+ real* getAllBladeVelocitiesYDevice(){ return this->bladeVelocitiesYDCurrentTimestep; };
+ real* getAllBladeVelocitiesZDevice(){ return this->bladeVelocitiesZDCurrentTimestep; };
+ real* getAllBladeForcesXDevice(){ return this->bladeForcesXDCurrentTimestep; };
+ real* getAllBladeForcesYDevice(){ return this->bladeForcesYDCurrentTimestep; };
+ real* getAllBladeForcesZDevice(){ return this->bladeForcesZDCurrentTimestep; };
+
+ real* getTurbineBladeRadiiDevice(uint turbine){ return &this->bladeRadiiD[turbine*numberOfBladeNodes]; };
+ real* getTurbineBladeCoordsXDevice(uint turbine){ return &this->bladeCoordsXDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeCoordsYDevice(uint turbine){ return &this->bladeCoordsYDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeCoordsZDevice(uint turbine){ return &this->bladeCoordsZDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeVelocitiesXDevice(uint turbine){ return &this->bladeVelocitiesXDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeVelocitiesYDevice(uint turbine){ return &this->bladeVelocitiesYDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeVelocitiesZDevice(uint turbine){ return &this->bladeVelocitiesZDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeForcesXDevice(uint turbine){ return &this->bladeForcesXDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeForcesYDevice(uint turbine){ return &this->bladeForcesYDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+ real* getTurbineBladeForcesZDevice(uint turbine){ return &this->bladeForcesZDCurrentTimestep[turbine*numberOfBladeNodes*numberOfBlades]; };
+
+ void setAllAzimuths(real* _azimuth);
+ void setAllOmegas(real* _omegas);
+ void setAllYaws(real* yaws);
+
+ void setTurbineAzimuth(uint turbine, real azimuth){ azimuthsH[turbine] = azimuth; };
+ void setTurbineYaw(uint turbine, real yaw){ yawsH[turbine] = yaw; };
+ void setTurbineOmega(uint turbine, real omega){ omegasH[turbine] = omega; };
+
+ void setAllBladeCoords(real* _bladeCoordsX, real* _bladeCoordsY, real* _bladeCoordsZ);
+ void setAllBladeVelocities(real* _bladeVelocitiesX, real* _bladeVelocitiesY, real* _bladeVelocitiesZ);
+ void setAllBladeForces(real* _bladeForcesX, real* _bladeForcesY, real* _bladeForcesZ);
+
+ void setTurbineBladeCoords(uint turbine, real* _bladeCoordsX, real* _bladeCoordsY, real* _bladeCoordsZ);
+ void setTurbineBladeVelocities(uint turbine, real* _bladeVelocitiesX, real* _bladeVelocitiesY, real* _bladeVelocitiesZ);
+ void setTurbineBladeForces(uint turbine, real* _bladeForcesX, real* _bladeForcesY, real* _bladeForcesZ);
+
+ virtual void calcBladeForces();
+
+private:
+ void initTurbineGeometries(CudaMemoryManager* cudaManager);
+ void initBoundingSpheres(Parameter* para, CudaMemoryManager* cudaManager);
+ void initBladeCoords(CudaMemoryManager* cudaManager);
+ void initBladeVelocities(CudaMemoryManager* cudaManager);
+ void initBladeForces(CudaMemoryManager* cudaManager);
+ void initBladeIndices(Parameter* para, CudaMemoryManager* cudaManager);
+
+ void calcForcesEllipticWing();
+ void rotateBlades(real angle, uint turbineID);
+
+ void writeBladeCoords(uint t);
+ void writeBladeForces(uint t);
+ void writeBladeVelocities(uint t);
+
+ void swapDeviceArrays();
+
+public:
+ real* bladeRadiiH;
+ real* bladeRadiiD;
+ real* bladeCoordsXH, * bladeCoordsYH, * bladeCoordsZH;
+ real* bladeCoordsXDPreviousTimestep, * bladeCoordsYDPreviousTimestep, * bladeCoordsZDPreviousTimestep;
+ real* bladeCoordsXDCurrentTimestep, * bladeCoordsYDCurrentTimestep, * bladeCoordsZDCurrentTimestep;
+ real* bladeVelocitiesXH, * bladeVelocitiesYH, * bladeVelocitiesZH;
+ real* bladeVelocitiesXDPreviousTimestep, * bladeVelocitiesYDPreviousTimestep, * bladeVelocitiesZDPreviousTimestep;
+ real* bladeVelocitiesXDCurrentTimestep, * bladeVelocitiesYDCurrentTimestep, * bladeVelocitiesZDCurrentTimestep;
+ real* bladeForcesXH, * bladeForcesYH, * bladeForcesZH;
+ real* bladeForcesXDPreviousTimestep, * bladeForcesYDPreviousTimestep, * bladeForcesZDPreviousTimestep;
+ real* bladeForcesXDCurrentTimestep, * bladeForcesYDCurrentTimestep, * bladeForcesZDCurrentTimestep;
+ uint* bladeIndicesH;
+ uint* bladeIndicesD;
+ uint* boundingSphereIndicesH;
+ uint* boundingSphereIndicesD;
+ real* turbinePosXH, *turbinePosYH, *turbinePosZH, *omegasH, *azimuthsH, *yawsH, *diametersH;
+ real* turbinePosXD, *turbinePosYD, *turbinePosZD, *omegasD, *azimuthsD, *yawsD, *diametersD;
+
+private:
+ std::vector<real> preInitPosX, preInitPosY, preInitPosZ, preInitDiameters, preInitOmegas, preInitAzimuths, preInitYaws;
+ std::vector<std::vector<real>> preInitBladeRadii;
+ const bool useHostArrays;
+ const real density;
+ real deltaT, deltaX;
+ const uint numberOfBladeNodes, numberOfBlades;
+ uint numberOfTurbines;
+ const real epsilon; // in m
+ const int level;
+ uint numberOfIndices;
+ uint numberOfNodes;
+ real forceRatio, factorGaussian, invEpsilonSqrd, invDeltaX;
+ int streamIndex;
+};
+
+#endif
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.cu b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.cu
deleted file mode 100644
index 71897bd21ea4fb299d3cc0ffa385506d4503f360..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.cu
+++ /dev/null
@@ -1,423 +0,0 @@
-#include "ActuatorLine.h"
-
-#include <cuda.h>
-#include <cuda_runtime.h>
-#include <helper_cuda.h>
-
-#include <cuda/CudaGrid.h>
-#include "VirtualFluids_GPU/GPU/GeometryUtils.h"
-
-#include "Parameter/Parameter.h"
-#include "DataStructureInitializer/GridProvider.h"
-#include "GPU/CudaMemoryManager.h"
-
-__host__ __device__ __inline__ uint calcNode(uint bladeNode, uint nBladeNodes, uint blade, uint nBlades)
-{
- return bladeNode+blade*nBladeNodes;
-}
-
-__host__ __device__ __inline__ void calcBladeAndBladeNode(uint node, uint& bladeNode, uint nBladeNodes, uint& blade, uint nBlades)
-{
- blade = node/nBladeNodes;
- bladeNode = node - blade*nBladeNodes;
-}
-
-__host__ __device__ __forceinline__ real distSqrd(real distX, real distY, real distZ)
-{
- return distX*distX+distY*distY+distZ*distZ;
-}
-
-__host__ __device__ __inline__ void rotateFromBladeToGlobal(
- real& bladeCoordX_BF, real& bladeCoordY_BF, real& bladeCoordZ_BF,
- real& bladeCoordX_GF, real& bladeCoordY_GF, real& bladeCoordZ_GF,
- real& azimuth, real& yaw)
-{
- real tmpX, tmpY, tmpZ;
-
- rotateAboutX3D(azimuth, bladeCoordX_BF, bladeCoordY_BF, bladeCoordZ_BF, tmpX, tmpY, tmpZ);
- rotateAboutZ3D(yaw, tmpX, tmpY, tmpZ, bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF);
-
-}
-
-__host__ __device__ __inline__ void rotateFromGlobalToBlade(
- real& bladeCoordX_BF, real& bladeCoordY_BF, real& bladeCoordZ_BF,
- real& bladeCoordX_GF, real& bladeCoordY_GF, real& bladeCoordZ_GF,
- real& azimuth, real& yaw)
-{
- real tmpX, tmpY, tmpZ;
-
- invRotateAboutZ3D(yaw, bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF, tmpX, tmpY, tmpZ);
- invRotateAboutX3D(azimuth, tmpX, tmpY, tmpZ, bladeCoordX_BF, bladeCoordY_BF, bladeCoordZ_BF);
-}
-
-__global__ void interpolateVelocities(real* gridCoordsX, real* gridCoordsY, real* gridCoordsZ,
- uint* neighborsX, uint* neighborsY, uint* neighborsZ, uint* neighborsWSB,
- real* vx, real* vy, real* vz,
- real* bladeCoordsX, real* bladeCoordsY, real* bladeCoordsZ,
- real* bladeVelocitiesX, real* bladeVelocitiesY, real* bladeVelocitiesZ,
- uint nBlades, uint nBladeNodes,
- real azimuth, real yaw, real omega,
- real turbPosX, real turbPosY, real turbPosZ,
- uint* bladeIndices, real velocityRatio, real invDeltaX)
-{
- const uint x = threadIdx.x;
- const uint y = blockIdx.x;
- const uint z = blockIdx.y;
-
- const uint nx = blockDim.x;
- const uint ny = gridDim.x;
-
- const uint node = nx*(ny*z + y) + x;
-
- uint bladeNode, blade;
-
- calcBladeAndBladeNode(node, bladeNode, nBladeNodes, blade, nBlades);
-
- if(node>=nBladeNodes*nBlades) return;
-
- real bladeCoordX_BF = bladeCoordsX[node];
- real bladeCoordY_BF = bladeCoordsY[node];
- real bladeCoordZ_BF = bladeCoordsZ[node];
-
- real bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF;
-
- real localAzimuth = azimuth+blade*c2Pi/nBlades;
-
- rotateFromBladeToGlobal(bladeCoordX_BF, bladeCoordY_BF, bladeCoordZ_BF,
- bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF,
- localAzimuth, yaw);
-
- bladeCoordX_GF += turbPosX;
- bladeCoordY_GF += turbPosY;
- bladeCoordZ_GF += turbPosZ;
-
- uint k, ke, kn, kt;
- uint kne, kte, ktn, ktne;
-
- k = findNearestCellBSW(bladeIndices[node],
- gridCoordsX, gridCoordsY, gridCoordsZ,
- bladeCoordX_GF, bladeCoordY_GF, bladeCoordZ_GF,
- neighborsX, neighborsY, neighborsZ, neighborsWSB);
-
- bladeIndices[node] = k;
-
- getNeighborIndicesOfBSW(k, ke, kn, kt, kne, kte, ktn, ktne, neighborsX, neighborsY, neighborsZ);
-
- real dW, dE, dN, dS, dT, dB;
-
- real distX = invDeltaX*(bladeCoordX_GF-gridCoordsX[k]);
- real distY = invDeltaX*(bladeCoordY_GF-gridCoordsY[k]);
- real distZ = invDeltaX*(bladeCoordZ_GF-gridCoordsZ[k]);
-
- getInterpolationWeights(dW, dE, dN, dS, dT, dB, distX, distY, distZ);
-
- real bladeVelX_GF = trilinearInterpolation(dW, dE, dN, dS, dT, dB, k, ke, kn, kt, kne, kte, ktn, ktne, vx)*velocityRatio;
- real bladeVelY_GF = trilinearInterpolation(dW, dE, dN, dS, dT, dB, k, ke, kn, kt, kne, kte, ktn, ktne, vy)*velocityRatio;
- real bladeVelZ_GF = trilinearInterpolation(dW, dE, dN, dS, dT, dB, k, ke, kn, kt, kne, kte, ktn, ktne, vz)*velocityRatio;
-
- real bladeVelX_BF, bladeVelY_BF, bladeVelZ_BF;
-
- rotateFromGlobalToBlade(bladeVelX_BF, bladeVelY_BF, bladeVelZ_BF,
- bladeVelX_GF, bladeVelY_GF, bladeVelZ_GF,
- localAzimuth, yaw);
-
- bladeVelocitiesX[node] = bladeVelX_BF;
- bladeVelocitiesY[node] = bladeVelY_BF+omega*bladeCoordZ_BF;
- bladeVelocitiesZ[node] = bladeVelZ_BF;
-}
-
-
-__global__ void applyBodyForces(real* gridCoordsX, real* gridCoordsY, real* gridCoordsZ,
- real* gridForcesX, real* gridForcesY, real* gridForcesZ,
- real* bladeCoordsX, real* bladeCoordsY, real* bladeCoordsZ,
- real* bladeForcesX, real* bladeForcesY,real* bladeForcesZ,
- uint nBlades, uint nBladeNodes,
- real azimuth, real yaw, real omega,
- real turbPosX, real turbPosY, real turbPosZ,
- uint* gridIndices, uint nIndices,
- real invEpsilonSqrd, real factorGaussian)
-{
- const uint x = threadIdx.x;
- const uint y = blockIdx.x;
- const uint z = blockIdx.y;
-
- const uint nx = blockDim.x;
- const uint ny = gridDim.x;
-
- const uint index = nx*(ny*z + y) + x;
-
- if(index>=nIndices) return;
-
- uint gridIndex = gridIndices[index];
-
- real gridCoordX_RF = gridCoordsX[gridIndex] - turbPosX;
- real gridCoordY_RF = gridCoordsY[gridIndex] - turbPosY;
- real gridCoordZ_RF = gridCoordsZ[gridIndex] - turbPosZ;
-
- real gridForceX_RF = c0o1;
- real gridForceY_RF = c0o1;
- real gridForceZ_RF = c0o1;
-
- real dAzimuth = c2Pi/nBlades;
-
- for( uint blade=0; blade<nBlades; blade++)
- {
- real localAzimuth = azimuth+blade*dAzimuth;
-
- real gridCoordX_BF, gridCoordY_BF, gridCoordZ_BF;
-
- rotateFromGlobalToBlade(gridCoordX_BF, gridCoordY_BF, gridCoordZ_BF,
- gridCoordX_RF, gridCoordY_RF, gridCoordZ_RF,
- localAzimuth, yaw);
-
- for( uint bladeNode=0; bladeNode<nBladeNodes; bladeNode++)
- {
- uint node = calcNode(bladeNode, nBladeNodes, blade, nBlades);
-
- real eta = factorGaussian*exp(-distSqrd(bladeCoordsX[node]-gridCoordX_BF, bladeCoordsY[node]-gridCoordY_BF, bladeCoordsZ[node]-gridCoordZ_BF)*invEpsilonSqrd);
-
- real forceX_RF, forceY_RF, forceZ_RF;
-
- rotateFromBladeToGlobal(bladeForcesX[node], bladeForcesY[node], bladeForcesZ[node],
- forceX_RF, forceY_RF, forceZ_RF,
- localAzimuth, yaw);
-
- gridForceX_RF += forceX_RF*eta;
- gridForceY_RF += forceY_RF*eta;
- gridForceZ_RF += forceZ_RF*eta;
- }
- }
-
- atomicAdd(&gridForcesX[gridIndex], gridForceX_RF);
- atomicAdd(&gridForcesY[gridIndex], gridForceY_RF);
- atomicAdd(&gridForcesZ[gridIndex], gridForceZ_RF);
-}
-
-
-void ActuatorLine::init(Parameter* para, GridProvider* gridProvider, CudaMemoryManager* cudaMemoryManager)
-{
- if(!para->getIsBodyForce()) throw std::runtime_error("try to allocate ActuatorLine but BodyForce is not set in Parameter.");
- this->initBladeRadii(cudaMemoryManager);
- this->initBladeCoords(cudaMemoryManager);
- this->initBladeIndices(para, cudaMemoryManager);
- this->initBladeVelocities(cudaMemoryManager);
- this->initBladeForces(cudaMemoryManager);
- this->initBoundingSphere(para, cudaMemoryManager);
-}
-
-
-void ActuatorLine::interact(Parameter* para, CudaMemoryManager* cudaMemoryManager, int level, unsigned int t)
-{
- if (level != this->level) return;
-
- cudaMemoryManager->cudaCopyBladeCoordsHtoD(this);
-
- vf::cuda::CudaGrid bladeGrid = vf::cuda::CudaGrid(para->getParH(level)->numberofthreads, this->nNodes);
-
- interpolateVelocities<<< bladeGrid.grid, bladeGrid.threads >>>(
- para->getParD(this->level)->coordinateX, para->getParD(this->level)->coordinateY, para->getParD(this->level)->coordinateZ,
- para->getParD(this->level)->neighborX, para->getParD(this->level)->neighborY, para->getParD(this->level)->neighborZ, para->getParD(this->level)->neighborInverse,
- para->getParD(this->level)->velocityX, para->getParD(this->level)->velocityY, para->getParD(this->level)->velocityZ,
- this->bladeCoordsXD, this->bladeCoordsYD, this->bladeCoordsZD,
- this->bladeVelocitiesXD, this->bladeVelocitiesYD, this->bladeVelocitiesZD,
- this->nBlades, this->nBladeNodes,
- this->azimuth, this->yaw, this->omega,
- this->turbinePosX, this->turbinePosY, this->turbinePosZ,
- this->bladeIndicesD, para->getVelocityRatio(), this->invDeltaX);
-
- cudaMemoryManager->cudaCopyBladeVelocitiesDtoH(this);
-
- this->calcBladeForces();
-
- cudaMemoryManager->cudaCopyBladeForcesHtoD(this);
-
- vf::cuda::CudaGrid sphereGrid = vf::cuda::CudaGrid(para->getParH(level)->numberofthreads, this->nIndices);
-
- applyBodyForces<<<sphereGrid.grid, sphereGrid.threads>>>(
- para->getParD(this->level)->coordinateX, para->getParD(this->level)->coordinateY, para->getParD(this->level)->coordinateZ,
- para->getParD(this->level)->forceX_SP, para->getParD(this->level)->forceY_SP, para->getParD(this->level)->forceZ_SP,
- this->bladeCoordsXD, this->bladeCoordsYD, this->bladeCoordsZD,
- this->bladeForcesXD, this->bladeForcesYD, this->bladeForcesZD,
- this->nBlades, this->nBladeNodes,
- this->azimuth, this->yaw, this->omega,
- this->turbinePosX, this->turbinePosY, this->turbinePosZ,
- this->boundingSphereIndicesD, this->nIndices,
- this->invEpsilonSqrd, this->factorGaussian);
-
- this->azimuth = fmod(this->azimuth+this->omega*this->deltaT,c2Pi);
-}
-
-
-void ActuatorLine::free(Parameter* para, CudaMemoryManager* cudaMemoryManager)
-{
- cudaMemoryManager->cudaFreeBladeRadii(this);
- cudaMemoryManager->cudaFreeBladeCoords(this);
- cudaMemoryManager->cudaFreeBladeVelocities(this);
- cudaMemoryManager->cudaFreeBladeForces(this);
- cudaMemoryManager->cudaFreeBladeIndices(this);
- cudaMemoryManager->cudaFreeSphereIndices(this);
-}
-
-
-void ActuatorLine::calcForcesEllipticWing()
-{
- uint node;
- real u_rel, v_rel, u_rel_sq;
- real phi;
- real Cl = c1o1;
- real Cd = c0o1;
- real c0 = c1o1;
-
- real c, Cn, Ct;
-
- for( uint blade=0; blade<this->nBlades; blade++)
- {
- for( uint bladeNode=0; bladeNode<this->nBladeNodes; bladeNode++)
- {
- node = calcNode(bladeNode, this->nBladeNodes, blade, this->nBlades);
-
- u_rel = this->bladeVelocitiesXH[node];
- v_rel = this->bladeVelocitiesYH[node];
- u_rel_sq = u_rel*u_rel+v_rel*v_rel;
- phi = atan2(u_rel, v_rel);
-
- real tmp = c4o1*this->bladeRadiiH[bladeNode]/this->diameter-c1o1;
- c = c0 * sqrt( c1o1- tmp*tmp );
- Cn = Cl*cos(phi)+Cd*sin(phi);
- Ct = Cl*sin(phi)-Cd*cos(phi);
-
- this->bladeForcesXH[node] = -c1o2*u_rel_sq*c*this->density*Cn;
- this->bladeForcesYH[node] = -c1o2*u_rel_sq*c*this->density*Ct;
- this->bladeForcesZH[node] = c0o1;
- }
- }
-}
-
-void ActuatorLine::calcBladeForces()
-{
- this->calcForcesEllipticWing();
-}
-
-void ActuatorLine::initBladeRadii(CudaMemoryManager* cudaMemoryManager)
-{
- cudaMemoryManager->cudaAllocBladeRadii(this);
-
- real dr = c1o2*this->diameter/this->nBladeNodes;
-
- for(uint node=0; node<this->nBladeNodes; node++)
- {
- this->bladeRadiiH[node] = dr*(node+1);
- }
- cudaMemoryManager->cudaCopyBladeRadiiHtoD(this);
-
- real dxOPiSqrtEps = pow(this->deltaX/(this->epsilon*sqrt(cPi)),c3o1);
- this->factorGaussian = dr*dxOPiSqrtEps/this->forceRatio;
-}
-
-void ActuatorLine::initBladeCoords(CudaMemoryManager* cudaMemoryManager)
-{
- cudaMemoryManager->cudaAllocBladeCoords(this);
-
- for(uint blade=0; blade<this->nBlades; blade++)
- {
- for(uint bladeNode=0; bladeNode<this->nBladeNodes; bladeNode++)
- {
- uint node = calcNode(bladeNode, this->nBladeNodes, blade, this->nBlades);
-
- this->bladeCoordsXH[node] = c0o1;
- this->bladeCoordsYH[node] = c0o1;
- this->bladeCoordsZH[node] = this->bladeRadiiH[bladeNode];
- }
- }
- cudaMemoryManager->cudaCopyBladeCoordsHtoD(this);
-}
-
-void ActuatorLine::initBladeVelocities(CudaMemoryManager* cudaMemoryManager)
-{
- cudaMemoryManager->cudaAllocBladeVelocities(this);
-
- for(uint node=0; node<this->nNodes; node++)
- {
- this->bladeVelocitiesXH[node] = c0o1;
- this->bladeVelocitiesYH[node] = c0o1;
- this->bladeVelocitiesZH[node] = c0o1;
- }
- cudaMemoryManager->cudaCopyBladeVelocitiesHtoD(this);
-}
-
-void ActuatorLine::initBladeForces(CudaMemoryManager* cudaMemoryManager)
-{
- cudaMemoryManager->cudaAllocBladeForces(this);
-
- for(uint node=0; node<this->nNodes; node++)
- {
- this->bladeForcesXH[node] = c0o1;
- this->bladeForcesYH[node] = c0o1;
- this->bladeForcesZH[node] = c0o1;
- }
- cudaMemoryManager->cudaCopyBladeForcesHtoD(this);
-}
-
-void ActuatorLine::initBladeIndices(Parameter* para, CudaMemoryManager* cudaMemoryManager)
-{
- cudaMemoryManager->cudaAllocBladeIndices(this);
-
- for(uint node=0; node<this->nNodes; node++)
-
- {
- this->bladeIndicesH[node] = 1;
- }
- cudaMemoryManager->cudaCopyBladeIndicesHtoD(this);
-}
-
-void ActuatorLine::initBoundingSphere(Parameter* para, CudaMemoryManager* cudaMemoryManager)
-{
- // Actuator line exists only on 1 level
- std::vector<int> nodesInSphere;
- real sphereRadius = c1o2*this->diameter+c4o1*this->epsilon;
- real sphereRadiusSqrd = sphereRadius*sphereRadius;
-
- for (uint j = 1; j <= para->getParH(this->level)->numberOfNodes; j++)
- {
- const real distX = para->getParH(this->level)->coordinateX[j]-this->turbinePosX;
- const real distY = para->getParH(this->level)->coordinateY[j]-this->turbinePosY;
- const real distZ = para->getParH(this->level)->coordinateZ[j]-this->turbinePosZ;
- if(distSqrd(distX,distY,distZ) < sphereRadiusSqrd) nodesInSphere.push_back(j);
- }
-
- this->nIndices = uint(nodesInSphere.size());
- cudaMemoryManager->cudaAllocSphereIndices(this);
- std::copy(nodesInSphere.begin(), nodesInSphere.end(), this->boundingSphereIndicesH);
- cudaMemoryManager->cudaCopySphereIndicesHtoD(this);
-}
-
-void ActuatorLine::setBladeCoords(real* _bladeCoordsX, real* _bladeCoordsY, real* _bladeCoordsZ)
-{
-
- for(uint node=0; node<this->nNodes; node++)
- {
- this->bladeCoordsXH[node] = _bladeCoordsX[node];
- this->bladeCoordsYH[node] = _bladeCoordsY[node];
- this->bladeCoordsZH[node] = _bladeCoordsZ[node];
- }
-}
-
-void ActuatorLine::setBladeVelocities(real* _bladeVelocitiesX, real* _bladeVelocitiesY, real* _bladeVelocitiesZ)
-{
- for(uint node=0; node<this->nNodes; node++)
- {
- this->bladeVelocitiesXH[node] = _bladeVelocitiesX[node];
- this->bladeVelocitiesYH[node] = _bladeVelocitiesY[node];
- this->bladeVelocitiesZH[node] = _bladeVelocitiesZ[node];
- }
-}
-
-void ActuatorLine::setBladeForces(real* _bladeForcesX, real* _bladeForcesY, real* _bladeForcesZ)
-{
- for(uint node=0; node<this->nNodes; node++)
- {
- this->bladeForcesXH[node] = _bladeForcesX[node];
- this->bladeForcesYH[node] = _bladeForcesY[node];
- this->bladeForcesZH[node] = _bladeForcesZ[node];
- }
-}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.h
deleted file mode 100644
index b44c89c5020eb206baa3bba1994b1e45f760c3bb..0000000000000000000000000000000000000000
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/ActuatorLine.h
+++ /dev/null
@@ -1,120 +0,0 @@
-#ifndef ActuatorLine_H
-#define ActuatorLine_H
-
-#include "PreCollisionInteractor.h"
-#include "PointerDefinitions.h"
-#include "VirtualFluids_GPU_export.h"
-#include "lbm/constants/NumericConstants.h"
-
-class Parameter;
-class GridProvider;
-
-using namespace vf::lbm::constant;
-class VIRTUALFLUIDS_GPU_EXPORT ActuatorLine : public PreCollisionInteractor
-{
-public:
- ActuatorLine(
- const uint _nBlades,
- const real _density,
- const uint _nBladeNodes,
- const real _epsilon,
- real _turbinePosX, real _turbinePosY, real _turbinePosZ,
- const real _diameter,
- int _level,
- const real _deltaT,
- const real _deltaX
- ) : nBlades(_nBlades),
- density(_density),
- nBladeNodes(_nBladeNodes),
- epsilon(_epsilon),
- turbinePosX(_turbinePosX), turbinePosY(_turbinePosY), turbinePosZ(_turbinePosZ),
- diameter(_diameter),
- level(_level),
- PreCollisionInteractor()
- {
- this->deltaT = _deltaT*exp2(-this->level);
- this->deltaX = _deltaX*exp2(-this->level);
- this->invDeltaX = c1o1/this->deltaX;
- this->forceRatio = this->density*pow(this->deltaX,4)*pow(this->deltaT,-2);
- this->invEpsilonSqrd = c1o1/(this->epsilon*this->epsilon);
- this->nNodes = this->nBladeNodes*this->nBlades;
- this->omega = c1o1;
- this->azimuth = c0o1;
- this->yaw = c0o1;
- };
-
- virtual ~ActuatorLine(){};
-
- void init(Parameter* para, GridProvider* gridProvider, CudaMemoryManager* cudaMemoryManager) override;
- void interact(Parameter* para, CudaMemoryManager* cudaMemoryManager, int level, uint t) override;
- void free(Parameter* para, CudaMemoryManager* cudaMemoryManager) override;
- void write(uint t);
-
- uint getNBladeNodes(){ return this->nBladeNodes; };
- uint getNBlades(){ return this->nBlades;};
- uint getNIndices(){ return this->nIndices; };
- uint getNNodes(){ return this->nNodes; };
- real getOmega(){ return this->omega; };
- real getAzimuth(){ return this->azimuth; };
- real getYaw(){ return this->yaw; };
- real getDensity(){ return this->density; };
- real getPositionX(){ return this->turbinePosX; };
- real getPositionY(){ return this->turbinePosY; };
- real getPositionZ(){ return this->turbinePosZ; };
- real* getBladeRadii(){ return this->bladeRadiiH; };
- real* getBladeCoordsX(){ return this->bladeCoordsXH; };
- real* getBladeCoordsY(){ return this->bladeCoordsYH; };
- real* getBladeCoordsZ(){ return this->bladeCoordsZH; };
- real* getBladeVelocitiesX(){ return this->bladeVelocitiesXH; };
- real* getBladeVelocitiesY(){ return this->bladeVelocitiesYH; };
- real* getBladeVelocitiesZ(){ return this->bladeVelocitiesZH; };
- real* getBladeForcesX(){ return this->bladeForcesXH; };
- real* getBladeForcesY(){ return this->bladeForcesYH; };
- real* getBladeForcesZ(){ return this->bladeForcesZH; };
-
- void setOmega(real _omega){ this->omega = _omega; };
- void setAzimuth(real _azimuth){ this->azimuth = _azimuth; };
- void setYaw(real _yaw){ this->yaw = _yaw; };
- void setBladeCoords(real* _bladeCoordsX, real* _bladeCoordsY, real* _bladeCoordsZ);
- void setBladeVelocities(real* _bladeVelocitiesX, real* _bladeVelocitiesY, real* _bladeVelocitiesZ);
- void setBladeForces(real* _bladeForcesX, real* _bladeForcesY, real* _bladeForcesZ);
- virtual void calcBladeForces();
-
-private:
- void initBoundingSphere(Parameter* para, CudaMemoryManager* cudaMemoryManager);
-
- void initBladeRadii(CudaMemoryManager* cudaMemoryManager);
- void initBladeCoords(CudaMemoryManager* cudaMemoryManager);
- void initBladeVelocities(CudaMemoryManager* cudaMemoryManager);
- void initBladeForces(CudaMemoryManager* cudaMemoryManager);
- void initBladeIndices(Parameter* para, CudaMemoryManager* cudaMemoryManager);
-
- void calcForcesEllipticWing();
-
-public:
- real* bladeRadiiH;
- real* bladeRadiiD;
- real* bladeCoordsXH, * bladeCoordsYH, * bladeCoordsZH;
- real* bladeCoordsXD, * bladeCoordsYD, * bladeCoordsZD;
- real* bladeVelocitiesXH, * bladeVelocitiesYH, * bladeVelocitiesZH;
- real* bladeVelocitiesXD, * bladeVelocitiesYD, * bladeVelocitiesZD;
- real* bladeForcesXH, * bladeForcesYH, * bladeForcesZH;
- real* bladeForcesXD, * bladeForcesYD, * bladeForcesZD;
- uint* bladeIndicesH;
- uint* bladeIndicesD;
- uint* boundingSphereIndicesH;
- uint* boundingSphereIndicesD;
-
-private:
- const real density;
- real turbinePosX, turbinePosY, turbinePosZ;
- real omega, azimuth, yaw, deltaT, deltaX, invDeltaX, forceRatio, factorGaussian, invEpsilonSqrd;
- const real diameter;
- const uint nBladeNodes;
- const uint nBlades;
- const real epsilon; // in m
- const int level;
- uint nIndices, nNodes;
-};
-
-#endif
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h
index a9b233f3035890c2617d3a00b639f995be6c218f..f9a87f613e7607301e59a7c1e67eb556418892e4 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PreCollisionInteractor.h
@@ -33,6 +33,7 @@ public:
virtual void init(Parameter *para, GridProvider *gridProvider, CudaMemoryManager *cudaMemoryManager) = 0;
virtual void interact(Parameter *para, CudaMemoryManager *cudaMemoryManager, int level, uint t) = 0;
virtual void free(Parameter *para, CudaMemoryManager *cudaMemoryManager) = 0;
+ virtual void getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider) = 0;
protected:
uint updateInterval;
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PrecursorWriter.cu b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PrecursorWriter.cu
new file mode 100644
index 0000000000000000000000000000000000000000..aacb1721ad90e9178b3c2ccf49a2aa9ec79d5504
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PrecursorWriter.cu
@@ -0,0 +1,321 @@
+#include "PrecursorWriter.h"
+#include "basics/writer/WbWriterVtkXmlImageBinary.h"
+
+#include <cuda.h>
+#include <cuda_runtime.h>
+#include <helper_cuda.h>
+#include <cuda/CudaGrid.h>
+#include "Kernel/Utilities/DistributionHelper.cuh"
+
+#include <Core/StringUtilities/StringUtil.h>
+
+#include "Parameter/Parameter.h"
+#include "DataStructureInitializer/GridProvider.h"
+#include "GPU/CudaMemoryManager.h"
+
+using namespace vf::lbm::dir;
+
+
+
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+//TODO check everything for multiple level
+void index1d(int& idx, int y, int z, int ny, int nz)
+{
+ idx = y+ny*z;
+}
+
+void index2d(int idx, int& y, int& z, int ny, int nz)
+{
+ z = idx/ny;
+ y = idx-ny*z;
+}
+
+__inline__ __host__ __device__ uint linearIdx(const uint component, const uint node, const uint timestep, const uint numberOfComponents, const uint numberOfNodes)
+{
+ return node+numberOfNodes*(component+numberOfComponents*timestep);
+}
+
+__inline__ __host__ __device__ uint linearIdx(const uint component, const uint node, const uint numberOfNodes)
+{
+ return node+component*numberOfNodes;
+}
+
+
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+__global__ void fillArrayVelocities(const uint numberOfPrecursorNodes,
+ uint* indices,
+ real *precursorData,
+ real *vx,
+ real *vy,
+ real *vz,
+ real velocityRatio)
+
+
+{
+ const uint node = vf::gpu::getNodeIndex();
+
+ if(node>=numberOfPrecursorNodes) return;
+
+ precursorData[linearIdx(0u, node, numberOfPrecursorNodes)] = vx[indices[node]]*velocityRatio;
+ precursorData[linearIdx(1u, node, numberOfPrecursorNodes)] = vy[indices[node]]*velocityRatio;
+ precursorData[linearIdx(2u, node, numberOfPrecursorNodes)] = vz[indices[node]]*velocityRatio;
+}
+
+
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+__global__ void fillArrayDistributions( uint numberOfPrecursorNodes,
+ uint* indices,
+ real* precursorData,
+ real* distributions,
+ uint* neighborX, uint* neighborY, uint* neighborZ,
+ bool isEvenTimestep,
+ unsigned long numberOfLBnodes)
+{
+ const uint node = vf::gpu::getNodeIndex();
+
+ if(node>=numberOfPrecursorNodes) return;
+
+ Distributions27 dist = vf::gpu::getDistributionReferences27(distributions, numberOfLBnodes, isEvenTimestep);
+
+ ////////////////////////////////////////////////////////////////////////////////
+ // ! - Set neighbor indices (necessary for indirect addressing)
+ uint k_000 = indices[node];
+ // uint k_M00 = neighborX[k_000];
+ uint k_0M0 = neighborY[k_000];
+ uint k_00M = neighborZ[k_000];
+ // uint k_MM0 = neighborY[k_M00];
+ // uint k_M0M = neighborZ[k_M00];
+ uint k_0MM = neighborZ[k_0M0];
+ // uint k_MMM = neighborZ[k_MM0];
+
+ ////////////////////////////////////////////////////////////////////////////////////
+ //! - Get local distributions in PX directions
+ //!
+ precursorData[linearIdx(PrecP00, node, numberOfPrecursorNodes)] = (dist.f[DIR_P00])[k_000];
+ precursorData[linearIdx(PrecPP0, node, numberOfPrecursorNodes)] = (dist.f[DIR_PP0])[k_000];
+ precursorData[linearIdx(PrecPM0, node, numberOfPrecursorNodes)] = (dist.f[DIR_PM0])[k_0M0];
+ precursorData[linearIdx(PrecP0P, node, numberOfPrecursorNodes)] = (dist.f[DIR_P0P])[k_000];
+ precursorData[linearIdx(PrecP0M, node, numberOfPrecursorNodes)] = (dist.f[DIR_P0M])[k_00M];
+ precursorData[linearIdx(PrecPPP, node, numberOfPrecursorNodes)] = (dist.f[DIR_PPP])[k_000];
+ precursorData[linearIdx(PrecPMP, node, numberOfPrecursorNodes)] = (dist.f[DIR_PMP])[k_0M0];
+ precursorData[linearIdx(PrecPPM, node, numberOfPrecursorNodes)] = (dist.f[DIR_PPM])[k_00M];
+ precursorData[linearIdx(PrecPMM, node, numberOfPrecursorNodes)] = (dist.f[DIR_PMM])[k_0MM];
+}
+
+
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+void PrecursorWriter::init(Parameter* para, GridProvider* gridProvider, CudaMemoryManager* cudaManager)
+{
+ VF_LOG_INFO("PrecursorWriter: Start initializing...");
+ VF_LOG_INFO("Writing yz-planes at x={}m every {}. timestep, starting at t={}", this->xPos, this->tSave, this->tStartOut);
+
+ precursorStructs.resize(para->getMaxLevel()+1);
+ for(int level=0; level<=para->getMaxLevel(); level++)
+ {
+
+ real dx = abs(para->getParH(level)->coordinateX[1]-para->getParH(level)->coordinateX[para->getParH(level)->neighborX[1]]);
+ int maxPoints = (int((yMax-yMin)/dx)+1)* (int((zMax-zMin)/dx)+1);
+
+ real lowestY, lowestZ, highestY, highestZ;
+
+ lowestY = para->getParH(level)->coordinateY[para->getParH(level)->numberOfNodes-1];
+ highestY = para->getParH(level)->coordinateY[1];
+
+ lowestZ = para->getParH(level)->coordinateZ[para->getParH(level)->numberOfNodes-1];
+ highestZ = para->getParH(level)->coordinateZ[1];
+
+ std::vector<uint> indicesOnGrid;
+ std::vector<int> indicesOnPlane;
+ std::vector<real> coordY, coordZ;
+
+ for(uint j=1; j<para->getParH(level)->numberOfNodes; j++ )
+ {
+ real pointCoordX = para->getParH(level)->coordinateX[j];
+ real pointCoordY = para->getParH(level)->coordinateY[j];
+ real pointCoordZ = para->getParH(level)->coordinateZ[j];
+ if( para->getParH(level)->typeOfGridNode[j] == GEO_FLUID &&
+ pointCoordX < (dx+xPos) && pointCoordX >= xPos &&
+ pointCoordY<=yMax && pointCoordY>=yMin &&
+ pointCoordZ<=zMax && pointCoordZ>=zMin)
+ {
+ highestY = max(highestY, pointCoordY);
+ highestZ = max(highestZ, pointCoordZ);
+
+ lowestY = min(lowestY, pointCoordY);
+ lowestZ = min(lowestZ, pointCoordZ);
+ indicesOnGrid.push_back(j);
+ coordY.push_back(pointCoordY);
+ coordZ.push_back(pointCoordZ);
+ }
+ }
+ if(indicesOnGrid.size()==0)
+ throw std::runtime_error("PrecursorWriter did not find any points on the grid");
+
+ int ny = int((highestY-lowestY)/dx)+1;
+ int nz = int((highestZ-lowestZ)/dx)+1;
+
+ for(uint i=0;i<indicesOnGrid.size(); i++)
+ {
+ int idxY = int((coordY[i]-lowestY)/dx);
+ int idxZ = int((coordZ[i]-lowestZ)/dx);
+ int idx;
+ index1d(idx, idxY, idxZ, ny, nz);
+ indicesOnPlane.push_back(idx);
+ }
+
+ precursorStructs[level] = SPtr<PrecursorStruct>(new PrecursorStruct);
+ precursorStructs[level]->numberOfPointsInBC = (uint)indicesOnGrid.size();
+ precursorStructs[level]->indicesOnPlane = (int*) malloc(precursorStructs[level]->numberOfPointsInBC*sizeof(int));
+ precursorStructs[level]->spacing = makeUbTuple(dx, dx, tSave*para->getTimeRatio()*pow(2,-level));
+ precursorStructs[level]->origin = makeUbTuple(lowestY, lowestZ);
+ precursorStructs[level]->extent = makeUbTuple(0, ny-1, 0, nz-1);
+ precursorStructs[level]->numberOfPointsInData = ny*nz;
+ precursorStructs[level]->numberOfTimestepsPerFile = min(para->getlimitOfNodesForVTK()/(ny*nz), maxtimestepsPerFile);
+ precursorStructs[level]->numberOfFilesWritten = 0;
+ precursorStructs[level]->numberOfTimestepsBuffered = 0;
+
+ switch (outputVariable)
+ {
+ case OutputVariable::Velocities:
+ precursorStructs[level]->numberOfQuantities = 3;
+ break;
+ case OutputVariable::Distributions:
+ precursorStructs[level]->numberOfQuantities = 9;
+ break;
+
+ default:
+ break;
+ }
+
+ cudaManager->cudaAllocPrecursorWriter(this, level);
+
+ std::copy(indicesOnGrid.begin(), indicesOnGrid.end(), precursorStructs[level]->indicesH);
+ std::copy(indicesOnPlane.begin(), indicesOnPlane.end(), precursorStructs[level]->indicesOnPlane);
+
+ cudaManager->cudaCopyPrecursorWriterIndicesHtoD(this, level);
+
+ VF_LOG_INFO("Found {} points in precursor plane on level {}", precursorStructs[level]->numberOfPointsInBC, level);
+ }
+ VF_LOG_INFO("PrecursorWriter: Done initializing.");
+}
+
+
+void PrecursorWriter::interact(Parameter* para, CudaMemoryManager* cudaManager, int level, uint t)
+{
+ uint t_level = para->getTimeStep(level, t, true);
+ uint tStartOut_level = tStartOut*pow(2, level);
+ uint tEnd_level = para->getTimestepEnd()*pow(2, level);
+
+ if(t_level>tStartOut_level && ((t_level-tStartOut_level) % tSave)==0)
+ {
+ vf::cuda::CudaGrid grid = vf::cuda::CudaGrid(para->getParH(level)->numberofthreads, precursorStructs[level]->numberOfPointsInBC);
+
+ if(this->outputVariable==OutputVariable::Velocities)
+ {
+ fillArrayVelocities<<<grid.grid, grid.threads>>>( precursorStructs[level]->numberOfPointsInBC, precursorStructs[level]->indicesD,
+ precursorStructs[level]->bufferD,
+ para->getParD(level)->velocityX, para->getParD(level)->velocityY, para->getParD(level)->velocityZ,
+ para->getVelocityRatio());
+ getLastCudaError("In PrecursorWriter::interact fillArrayVelocities execution failed");
+ }
+ else if(this->outputVariable==OutputVariable::Distributions)
+ {
+ fillArrayDistributions<<<grid.grid, grid.threads>>>(precursorStructs[level]->numberOfPointsInBC, precursorStructs[level]->indicesD,
+ precursorStructs[level]->bufferD,
+ para->getParD(level)->distributions.f[0],
+ para->getParD(level)->neighborX, para->getParD(level)->neighborY, para->getParD(level)->neighborZ,
+ para->getEvenOrOdd(level), para->getParD(level)->numberOfNodes);
+ getLastCudaError("In PrecursorWriter::interact fillArrayDistributions execution failed");
+ }
+ cudaManager->cudaCopyPrecursorWriterOutputVariablesDtoH(this, level);
+
+ // switch device buffer and data pointer so precursor data is gathered in buffer and copied from bufferD to bufferH
+ real *tmp = precursorStructs[level]->bufferD;
+ precursorStructs[level]->bufferD = precursorStructs[level]->dataD;
+ precursorStructs[level]->dataD = tmp;
+
+ precursorStructs[level]->numberOfTimestepsBuffered++;
+
+ if(precursorStructs[level]->numberOfTimestepsBuffered >= precursorStructs[level]->numberOfTimestepsPerFile || t == para->getTimestepEnd())
+ {
+ // switch host buffer and data pointer so precursor data is copied in buffer and written from data
+
+ tmp = precursorStructs[level]->bufferH;
+ precursorStructs[level]->bufferH = precursorStructs[level]->dataH;
+ precursorStructs[level]->dataH = tmp;
+
+ writeFuture.wait();
+ writeFuture = std::async(std::launch::async, [this](Parameter* para, uint level, uint timesteps){ this->write(para, level, timesteps); }, para, level, precursorStructs[level]->numberOfTimestepsBuffered);
+ precursorStructs[level]->numberOfTimestepsBuffered = 0;
+ }
+ }
+}
+
+
+void PrecursorWriter::free(Parameter* para, CudaMemoryManager* cudaManager)
+{
+ writeFuture.wait();
+ for(int level=0; level<=para->getMaxLevel(); level++)
+ {
+ if(getPrecursorStruct(level)->numberOfTimestepsBuffered>0)
+ write(para, level, getPrecursorStruct(level)->numberOfTimestepsBuffered);
+
+ cudaManager->cudaFreePrecursorWriter(this, level);
+ }
+}
+
+
+void PrecursorWriter::write(Parameter* para, int level, uint numberOfTimestepsBuffered)
+{
+ std::string fname = this->makeFileName(fileName, level, para->getMyProcessID(), precursorStructs[level]->numberOfFilesWritten) + getWriter()->getFileExtension();
+ std::string wholeName = outputPath + "/" + fname;
+
+ uint numberOfPointsInData = precursorStructs[level]->numberOfPointsInData;
+
+ int startTime = precursorStructs[level]->numberOfFilesWritten*precursorStructs[level]->numberOfTimestepsPerFile;
+
+ UbTupleInt6 extent = makeUbTuple( val<1>(precursorStructs[level]->extent), val<2>(precursorStructs[level]->extent),
+ val<3>(precursorStructs[level]->extent), val<4>(precursorStructs[level]->extent),
+ startTime, startTime+(int)numberOfTimestepsBuffered-1);
+
+ UbTupleFloat3 origin = makeUbTuple( val<1>(precursorStructs[level]->origin), val<2>(precursorStructs[level]->origin), 0.f);
+
+ std::vector<std::vector<double>> nodedata;
+
+ for(uint quant=0; quant<precursorStructs[level]->numberOfQuantities; quant++)
+ {
+ std::vector<double> doubleArr(numberOfPointsInData*numberOfTimestepsBuffered, NAN);
+ for( uint timestep=0; timestep<numberOfTimestepsBuffered; timestep++)
+ {
+ for (uint pos=0; pos < precursorStructs[level]->numberOfPointsInBC; pos++)
+ {
+ int indexOnPlane = precursorStructs[level]->indicesOnPlane[pos]+timestep*numberOfPointsInData;
+ doubleArr[indexOnPlane] = double(precursorStructs[level]->dataH[linearIdx(quant, pos, timestep, precursorStructs[level]->numberOfQuantities, precursorStructs[level]->numberOfPointsInBC)]);
+ }
+ }
+ nodedata.push_back(doubleArr);
+ }
+
+ std::vector<std::vector<double>> celldata;
+ getWriter()->writeData(wholeName, nodedatanames, celldatanames, nodedata, celldata, extent, origin, precursorStructs[level]->spacing, extent, this->writePrecision);
+ precursorStructs[level]->numberOfFilesWritten++;
+}
+
+std::string PrecursorWriter::makeFileName(std::string fileName, int level, int id, uint numberOfFilesWritten)
+{
+ return fileName + "_lev_" + StringUtil::toString<int>(level)
+ + "_ID_" + StringUtil::toString<int>(id)
+ + "_File_" + StringUtil::toString<int>(numberOfFilesWritten);
+}
+
+void PrecursorWriter::getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider)
+{
+ for(uint level=0; level<(uint)para->getMaxLevel(); level++)
+ {
+ if(outputVariable==OutputVariable::Velocities)
+ {
+ std::vector<uint> indices(precursorStructs[level]->indicesH, precursorStructs[level]->indicesH+precursorStructs[level]->numberOfPointsInBC);
+ gridProvider->tagFluidNodeIndices(indices, CollisionTemplate::WriteMacroVars, level);
+ }
+ }
+}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PrecursorWriter.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PrecursorWriter.h
new file mode 100644
index 0000000000000000000000000000000000000000..3959fb2d8f5e719084d2d55e6f6ef84076383b45
--- /dev/null
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/PrecursorWriter.h
@@ -0,0 +1,161 @@
+//=======================================================================================
+// ____ ____ __ ______ __________ __ __ __ __
+// \ \ | | | | | _ \ |___ ___| | | | | / \ | |
+// \ \ | | | | | |_) | | | | | | | / \ | |
+// \ \ | | | | | _ / | | | | | | / /\ \ | |
+// \ \ | | | | | | \ \ | | | \__/ | / ____ \ | |____
+// \ \ | | |__| |__| \__\ |__| \________/ /__/ \__\ |_______|
+// \ \ | | ________________________________________________________________
+// \ \ | | | ______________________________________________________________|
+// \ \| | | | __ __ __ __ ______ _______
+// \ | | |_____ | | | | | | | | | _ \ / _____)
+// \ | | _____| | | | | | | | | | | \ \ \_______
+// \ | | | | |_____ | \_/ | | | | |_/ / _____ |
+// \ _____| |__| |________| \_______/ |__| |______/ (_______/
+//
+// This file is part of VirtualFluids. VirtualFluids is free software: you can
+// redistribute it and/or modify it under the terms of the GNU General Public
+// License as published by the Free Software Foundation, either version 3 of
+// the License, or (at your option) any later version.
+//
+// VirtualFluids is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+// for more details.
+//
+// You should have received a copy of the GNU General Public License along
+// with VirtualFluids (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
+//
+//! \file PrecursorWriter.h
+//! \author Henry Korb, Henrik Asmuth
+//! \date 05/12/2022
+//! \brief Probe writing planes of data to be used as inflow data in successor simulation using PrecursorBC
+//!
+//! The probe writes out yz-planes at a specifc x position ( \param xPos ) of either velocity or distributions
+//! that can be read by PrecursorBC as inflow data.
+//=======================================================================================
+
+
+#ifndef PRECURSORPROBE_H_
+#define PRECURSORPROBE_H_
+
+#include "PreCollisionInteractor.h"
+#include "WbWriterVtkXmlImageBinary.h"
+#include "LBM/LB.h"
+#include <string>
+#include <vector>
+#include <future>
+#include "PointerDefinitions.h"
+#include "Logger.h"
+
+class Parameter;
+class CudaMemoryManager;
+class GridProvider;
+
+enum class OutputVariable {
+ //! - Velocities
+ Velocities,
+ //! - Distributions
+ Distributions
+};
+
+static constexpr uint PrecP00 = 0;
+static constexpr uint PrecPP0 = 1;
+static constexpr uint PrecPM0 = 2;
+static constexpr uint PrecP0P = 3;
+static constexpr uint PrecP0M = 4;
+static constexpr uint PrecPPP = 5;
+static constexpr uint PrecPMP = 6;
+static constexpr uint PrecPPM = 7;
+static constexpr uint PrecPMM = 8;
+
+struct PrecursorStruct
+{
+ uint numberOfPointsInBC, numberOfPointsInData, numberOfTimestepsPerFile, numberOfFilesWritten, numberOfTimestepsBuffered;
+ uint *indicesH, *indicesD;
+ real *dataH, *dataD;
+ real *bufferH, *bufferD;
+ uint numberOfQuantities;
+ UbTupleInt4 extent;
+ UbTupleFloat2 origin;
+ UbTupleFloat3 spacing;
+ int* indicesOnPlane;
+ cudaStream_t stream;
+};
+
+class PrecursorWriter : public PreCollisionInteractor
+{
+public:
+ PrecursorWriter(
+ const std::string _fileName,
+ const std::string _outputPath,
+ real _xPos,
+ real _yMin, real _yMax,
+ real _zMin, real _zMax,
+ uint _tStartOut,
+ uint _tSave,
+ OutputVariable _outputVariable,
+ uint _maxTimestepsPerFile=uint(1e4)
+ ):
+ fileName(_fileName),
+ outputPath(_outputPath),
+ xPos(_xPos),
+ yMin(_yMin),
+ yMax(_yMax),
+ zMin(_zMin),
+ zMax(_zMax),
+ tStartOut(_tStartOut),
+ tSave(_tSave),
+ outputVariable(_outputVariable),
+ maxtimestepsPerFile(_maxTimestepsPerFile)
+ {
+ nodedatanames = determineNodeDataNames();
+ writeFuture = std::async([](){});
+ };
+
+ void init(Parameter* para, GridProvider* gridProvider, CudaMemoryManager* cudaManager) override;
+ void interact(Parameter* para, CudaMemoryManager* cudaManager, int level, uint t) override;
+ void free(Parameter* para, CudaMemoryManager* cudaManager) override;
+ void getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider) override;
+
+ OutputVariable getOutputVariable(){ return this->outputVariable; }
+
+ SPtr<PrecursorStruct> getPrecursorStruct(int level){return precursorStructs[level];}
+ static std::string makeFileName(std::string fileName, int level, int id, uint part);
+
+ void setWritePrecision(uint _writePrecision){ this->writePrecision=_writePrecision;}
+
+private:
+ WbWriterVtkXmlImageBinary* getWriter(){ return WbWriterVtkXmlImageBinary::getInstance(); };
+ void write(Parameter* para, int level, uint numberOfTimestepsBuffered);
+
+ std::vector<std::string> determineNodeDataNames()
+ {
+ switch (outputVariable)
+ {
+ case OutputVariable::Velocities:
+ return {"vx", "vy", "vz"};
+ break;
+ case OutputVariable::Distributions:
+ return {"fP00", "fPP0", "fPM0", "fP0P", "fP0M", "fPPP", "fPMP", "fPPM", "fPMM"};
+ break;
+
+ default:
+ throw std::runtime_error("Invalid OutputVariable for PrecursorWriter");
+ break;
+ }
+ }
+
+private:
+ std::vector<SPtr<PrecursorStruct>> precursorStructs;
+ std::string fileName, outputPath;
+ std::vector<std::string> nodedatanames;
+ std::vector<std::string> celldatanames;
+ uint tStartOut, tSave, maxtimestepsPerFile;
+ real xPos, yMin, yMax, zMin, zMax;
+ OutputVariable outputVariable;
+ std::future<void> writeFuture;
+ uint writePrecision = 8;
+};
+
+#endif //PRECURSORPROBE_H_
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.cu b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.cu
index f5b520acfad74f6787e9e657fce3ccdceed9d539..4f4affd9fffdfd2d7a91a14051ff49b354473b6a 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.cu
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.cu
@@ -15,6 +15,7 @@
#include "Parameter/Parameter.h"
#include "DataStructureInitializer/GridProvider.h"
#include "GPU/CudaMemoryManager.h"
+#include "GPU/GPU_Interface.h"
#include <algorithm>
@@ -268,6 +269,15 @@ void PlanarAverageProbe::findPoints(Parameter* para, GridProvider* gridProvider,
void PlanarAverageProbe::calculateQuantities(SPtr<ProbeStruct> probeStruct, Parameter* para, uint t_level, int level)
{
+ // Compute macroscopic variables in entire domain
+ CalcMacCompSP27(para->getParD(level)->velocityX, para->getParD(level)->velocityY, para->getParD(level)->velocityZ,
+ para->getParD(level)->rho, para->getParD(level)->pressure, para->getParD(level)->typeOfGridNode,
+ para->getParD(level)->neighborX, para->getParD(level)->neighborY,
+ para->getParD(level)->neighborZ, para->getParD(level)->numberOfNodes,
+ para->getParD(level)->numberofthreads, para->getParD(level)->distributions.f[0],
+ para->getParD(level)->isEvenTimestep);
+ getLastCudaError("In PlanarAverageProbe Kernel CalcMacSP27 execution failed");
+
// Definition of normal and inplane directions for moveIndices kernels
uint *neighborNormal, *neighborInplane1, *neighborInplane2;
if( this->planeNormal == 'x' )
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.h
index d11f8e76e4d13113b201af5494b7d0cfcfe18353..3d3533f74501e776f9150c83c9d9101a0be7ecbc 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.h
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlanarAverageProbe.h
@@ -74,6 +74,7 @@ public:
planeNormal(_planeNormal)
{
+ if (_tStartTmpAvg<_tStartAvg) throw std::runtime_error("Probe: tStartTmpAvg must be larger than tStartAvg!");
if(!(_planeNormal == 'x' || _planeNormal == 'y' || _planeNormal == 'z'))
throw std::runtime_error("PlanarAverageProbe: planeNormal must be 'x', 'y' or 'z'!");
}
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.cu b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.cu
index 7d1c0205219737e4b28acbb1a893a0a6071ae9de..8e0bf1faf87a251d41f1b5deeda698c75d9706c6 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.cu
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.cu
@@ -106,4 +106,14 @@ void PlaneProbe::calculateQuantities(SPtr<ProbeStruct> probeStruct, Parameter* p
para->getParD(level)->velocityX, para->getParD(level)->velocityY, para->getParD(level)->velocityZ, para->getParD(level)->rho,
para->getParD(level)->neighborX, para->getParD(level)->neighborY, para->getParD(level)->neighborZ,
probeStruct->quantitiesD, probeStruct->arrayOffsetsD, probeStruct->quantitiesArrayD);
+}
+
+void PlaneProbe::getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider)
+{
+ for(int level=0; level<=para->getMaxLevel(); level++)
+ {
+ SPtr<ProbeStruct> probeStruct = this->getProbeStruct(level);
+ std::vector<uint> probeIndices( probeStruct->pointIndicesH, probeStruct->pointIndicesH+probeStruct->nIndices);
+ gridProvider->tagFluidNodeIndices( probeIndices, CollisionTemplate::WriteMacroVars, level);
+ }
}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.h
index 3440c01020f9b3505be7148024e47373b76648ff..180169707a6d7f3f7975f6a2bc4009f7c0aba527 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.h
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PlaneProbe.h
@@ -54,7 +54,7 @@ public:
): Probe(_probeName,
_outputPath,
_tStartAvg,
- 0,
+ _tStartAvg+1,
_tAvg,
_tStartOut,
_tOut,
@@ -72,6 +72,8 @@ public:
this->deltaZ = _deltaZ;
}
+ void getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider) override;
+
private:
bool isAvailableStatistic(Statistic _variable) override;
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.cu b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.cu
index e78a98f02ac2093fc46b4daa4a2485ed1395275b..37beeed05e498028356cd89b416e5269c070de81 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.cu
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.cu
@@ -140,4 +140,14 @@ void PointProbe::addProbePointsFromXNormalPlane(real pos_x, real pos0_y, real po
}
printf("Added %u points \n", n_y*n_z);
+}
+
+void PointProbe::getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider)
+{
+ for(int level=0; level<=para->getMaxLevel(); level++)
+ {
+ SPtr<ProbeStruct> probeStruct = this->getProbeStruct(level);
+ std::vector<uint> probeIndices( probeStruct->pointIndicesH, probeStruct->pointIndicesH+probeStruct->nIndices);
+ gridProvider->tagFluidNodeIndices( probeIndices, CollisionTemplate::WriteMacroVars, level);
+ }
}
\ No newline at end of file
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.h
index 6a6fbe76f089acfafc22672dd3e9d71bd193a3b3..08c359705f03b20fbd3276fe209b6ff4d782a5e5 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.h
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/PointProbe.h
@@ -64,6 +64,7 @@ public:
void addProbePointsFromList(std::vector<real>& _pointCoordsX, std::vector<real>& _pointCoordsY, std::vector<real>& _pointCoordsZ);
void addProbePointsFromXNormalPlane(real pos_x, real pos0_y, real pos0_z, real pos1_y, real pos1_z, uint n_y, uint n_z);
+ void getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider) override;
private:
bool isAvailableStatistic(Statistic _variable) override;
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.cu b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.cu
index cc027b07bded01455437e65e08ccdcd51bcf7dc0..03c18f5a9a2133bec244053113209abc70469a2a 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.cu
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.cu
@@ -187,7 +187,7 @@ void Probe::init(Parameter* para, GridProvider* gridProvider, CudaMemoryManager*
this->velocityRatio = std::bind(&Parameter::getScaledVelocityRatio, para, _1);
this->densityRatio = std::bind(&Parameter::getScaledDensityRatio, para, _1);
this->forceRatio = std::bind(&Parameter::getScaledForceRatio, para, _1);
- this->stressRatio = std::bind(&Parameter::getScaledPressureRatio, para, _1);
+ this->stressRatio = std::bind(&Parameter::getScaledStressRatio, para, _1);
this->viscosityRatio = std::bind(&Parameter::getScaledViscosityRatio, para, _1);
this->nondimensional = std::bind(&Probe::getNondimensionalConversionFactor, this, _1);
@@ -315,6 +315,12 @@ void Probe::free(Parameter* para, CudaMemoryManager* cudaMemoryManager)
}
}
+void Probe::getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider)
+{
+ // Do nothing
+};
+
+
void Probe::addStatistic(Statistic variable)
{
if (!this->isAvailableStatistic(variable)) throw std::runtime_error("Probe::addStatistic(): Statistic not available for this probe type!");
@@ -329,6 +335,22 @@ void Probe::addStatistic(Statistic variable)
}
}
+std::string Probe::makeParallelFileName(int id, int t)
+{
+ return this->probeName + "_bin_ID_" + StringUtil::toString<int>(id)
+ + "_t_" + StringUtil::toString<int>(t)
+ + ".vtk";
+}
+
+std::string Probe::makeGridFileName(int level, int id, int t, uint part)
+{
+ return this->probeName + "_bin_lev_" + StringUtil::toString<int>(level)
+ + "_ID_" + StringUtil::toString<int>(id)
+ + "_Part_" + StringUtil::toString<int>(part)
+ + "_t_" + StringUtil::toString<int>(t)
+ + ".vtk";
+}
+
void Probe::addAllAvailableStatistics()
{
for( int var=0; var < int(Statistic::LAST); var++)
@@ -347,119 +369,76 @@ void Probe::write(Parameter* para, int level, int t)
std::vector<std::string> fnames;
for (uint i = 1; i <= numberOfParts; i++)
{
- std::string fname = this->probeName + "_bin_lev_" + StringUtil::toString<int>(level)
- + "_ID_" + StringUtil::toString<int>(para->getMyProcessID())
- + "_Part_" + StringUtil::toString<int>(i);
- if(!this->outputTimeSeries) fname += "_t_" + StringUtil::toString<int>(t_write);
- fname += ".vtk";
- fnames.push_back(fname);
- this->fileNamesForCollectionFile.push_back(fname);
+ this->writeGridFile(para, level, t_write, i);
}
- this->writeGridFiles(para, level, fnames, t);
-
- if(level == 0 && !this->outputTimeSeries) this->writeCollectionFile(para, t);
+ if(level == 0&& !this->outputTimeSeries) this->writeParallelFile(para, t);
}
-void Probe::writeCollectionFile(Parameter* para, int t)
+void Probe::writeParallelFile(Parameter* para, int t)
{
int t_write = this->fileNameLU ? t: t/this->tOut;
- std::string filename = this->probeName + "_bin_ID_" + StringUtil::toString<int>(para->getMyProcessID())
- + "_t_" + StringUtil::toString<int>(t_write)
- + ".vtk";
-
- std::ofstream file;
-
- file.open(this->outputPath + "/" + filename + ".pvtu" );
-
- //////////////////////////////////////////////////////////////////////////
-
- file << "<VTKFile type=\"PUnstructuredGrid\" version=\"1.0\" byte_order=\"LittleEndian\" header_type=\"UInt64\">" << std::endl;
- file << " <PUnstructuredGrid GhostLevel=\"1\">" << std::endl;
-
- file << " <PPointData>" << std::endl;
-
- for(std::string varName: this->getVarNames()) //TODO
- {
- file << " <DataArray type=\"Float64\" Name=\""<< varName << "\" /> " << std::endl;
- }
- file << " </PPointData>" << std::endl;
-
- file << " <PPoints>" << std::endl;
- file << " <PDataArray type=\"Float32\" Name=\"Points\" NumberOfComponents=\"3\"/>" << std::endl;
- file << " </PPoints>" << std::endl;
-
- for( auto& fname : this->fileNamesForCollectionFile )
- {
- const auto filenameWithoutPath=fname.substr( fname.find_last_of('/') + 1 );
- file << " <Piece Source=\"" << filenameWithoutPath << ".bin.vtu\"/>" << std::endl;
- }
-
- file << " </PUnstructuredGrid>" << std::endl;
- file << "</VTKFile>" << std::endl;
+ std::string filename = this->outputPath + "/" + this->makeParallelFileName(para->getMyProcessID(), t_write);
- //////////////////////////////////////////////////////////////////////////
+ std::vector<std::string> nodedatanames = this->getVarNames();
+ std::vector<std::string> cellNames;
- file.close();
+ getWriter()->writeParallelFile(filename, fileNamesForCollectionFile, nodedatanames, cellNames);
this->fileNamesForCollectionFile.clear();
}
-void Probe::writeGridFiles(Parameter* para, int level, std::vector<std::string>& fnames, int t)
+void Probe::writeGridFile(Parameter* para, int level, int t, uint part)
{
+ std::string fname = this->outputPath + "/" + this->makeGridFileName(level, para->getMyProcessID(), t, part);
+
std::vector< UbTupleFloat3 > nodes;
std::vector< std::string > nodedatanames = this->getVarNames();
- uint startpos = 0;
- uint endpos = 0;
- uint sizeOfNodes = 0;
std::vector< std::vector< double > > nodedata(nodedatanames.size());
SPtr<ProbeStruct> probeStruct = this->getProbeStruct(level);
- for (uint part = 0; part < fnames.size(); part++)
- {
- startpos = part * para->getlimitOfNodesForVTK();
- uint nDataPoints = this->outputTimeSeries? this->tProbe: probeStruct->nPoints;
- sizeOfNodes = min(para->getlimitOfNodesForVTK(), nDataPoints - startpos);
- endpos = startpos + sizeOfNodes;
+ uint startpos = (part-1) * para->getlimitOfNodesForVTK();
+ uint sizeOfNodes = min(para->getlimitOfNodesForVTK(), probeStruct->nPoints - startpos);
+ uint endpos = startpos + sizeOfNodes;
- //////////////////////////////////////////////////////////////////////////
- nodes.resize(sizeOfNodes);
+ //////////////////////////////////////////////////////////////////////////
+ nodes.resize(sizeOfNodes);
- for (uint pos = startpos; pos < endpos; pos++)
- {
- nodes[pos-startpos] = makeUbTuple( float(probeStruct->pointCoordsX[pos]),
- float(probeStruct->pointCoordsY[pos]),
- float(probeStruct->pointCoordsZ[pos]));
- }
+ for (uint pos = startpos; pos < endpos; pos++)
+ {
+ nodes[pos-startpos] = makeUbTuple( float(probeStruct->pointCoordsX[pos]),
+ float(probeStruct->pointCoordsY[pos]),
+ float(probeStruct->pointCoordsZ[pos]));
+ }
- for( auto it=nodedata.begin(); it!=nodedata.end(); it++) it->resize(sizeOfNodes);
+ for( auto it=nodedata.begin(); it!=nodedata.end(); it++) it->resize(sizeOfNodes);
- for( int var=0; var < int(Statistic::LAST); var++){
- if(this->quantities[var])
- {
- Statistic statistic = static_cast<Statistic>(var);
- real coeff;
+ for( int var=0; var < int(Statistic::LAST); var++){
+ if(this->quantities[var])
+ {
+ Statistic statistic = static_cast<Statistic>(var);
+ real coeff;
- std::vector<PostProcessingVariable> postProcessingVariables = this->getPostProcessingVariables(statistic);
- uint n_arrs = uint(postProcessingVariables.size());
+ std::vector<PostProcessingVariable> postProcessingVariables = this->getPostProcessingVariables(statistic);
+ uint n_arrs = uint(postProcessingVariables.size());
- uint arrOff = probeStruct->arrayOffsetsH[var];
- uint arrLen = probeStruct->nPoints;
+ uint arrOff = probeStruct->arrayOffsetsH[var];
+ uint arrLen = probeStruct->nPoints;
+
+ for(uint arr=0; arr<n_arrs; arr++)
+ {
+ coeff = postProcessingVariables[arr].conversionFactor(level);
- for(uint arr=0; arr<n_arrs; arr++)
+ for (uint pos = startpos; pos < endpos; pos++)
{
- coeff = postProcessingVariables[arr].conversionFactor(level);
-
- for (uint pos = startpos; pos < endpos; pos++)
- {
- nodedata[arrOff+arr][pos-startpos] = double(probeStruct->quantitiesArrayH[(arrOff+arr)*arrLen+pos]*coeff);
- }
+ nodedata[arrOff+arr][pos-startpos] = double(probeStruct->quantitiesArrayH[(arrOff+arr)*arrLen+pos]*coeff);
}
}
}
- WbWriterVtkXmlBinary::getInstance()->writeNodesWithNodeData(this->outputPath + "/" + fnames[part], nodes, nodedatanames, nodedata);
}
+ std::string fullName = getWriter()->writeNodesWithNodeData(fname, nodes, nodedatanames, nodedata);
+ this->fileNamesForCollectionFile.push_back(fullName.substr(fullName.find_last_of('/') + 1));
}
std::vector<std::string> Probe::getVarNames()
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.h
index 9cb0bd43e27fb7a28cae9c363ce245fbd9cc5677..aaf294e87d23c64707a16692b9337d6e9ff9c896 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.h
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/Probe.h
@@ -49,6 +49,7 @@
#include "PreCollisionInteractor/PreCollisionInteractor.h"
#include "PointerDefinitions.h"
+#include "WbWriterVtkXmlBinary.h"
//=======================================================================================
//! \note How to add new Statistics
@@ -153,12 +154,12 @@ public:
PreCollisionInteractor()
{
if (_tStartOut<_tStartAvg) throw std::runtime_error("Probe: tStartOut must be larger than tStartAvg!");
- if (_tStartTmpAvg<_tStartAvg) throw std::runtime_error("Probe: tStartTmpAvg must be larger than tStartAvg!");
}
void init(Parameter* para, GridProvider* gridProvider, CudaMemoryManager* cudaMemoryManager) override;
void interact(Parameter* para, CudaMemoryManager* cudaMemoryManager, int level, uint t) override;
void free(Parameter* para, CudaMemoryManager* cudaMemoryManager) override;
+ virtual void getTaggedFluidNodes(Parameter *para, GridProvider* gridProvider) override;
SPtr<ProbeStruct> getProbeStruct(int level){ return this->probeParams[level]; }
@@ -171,6 +172,8 @@ public:
void setFileNameToNOut(){this->fileNameLU = false;}
void setTStartTmpAveraging(uint _tStartTmpAveraging){this->tStartTmpAveraging = _tStartTmpAveraging;}
+protected:
+ virtual WbWriterVtkXmlBinary* getWriter(){ return WbWriterVtkXmlBinary::getInstance(); };
real getNondimensionalConversionFactor(int level);
private:
@@ -188,12 +191,15 @@ private:
int level);
virtual void calculateQuantities(SPtr<ProbeStruct> probeStruct, Parameter* para, uint t, int level) = 0;
- void write(Parameter* para, int level, int t);
- void writeCollectionFile(Parameter* para, int t);
- void writeGridFiles(Parameter* para, int level, std::vector<std::string >& fnames, int t);
+ virtual void write(Parameter* para, int level, int t);
+ virtual void writeParallelFile(Parameter* para, int t);
+ virtual void writeGridFile(Parameter* para, int level, int t, uint part);
+
std::vector<std::string> getVarNames();
-
-private:
+ std::string makeGridFileName(int level, int id, int t, uint part);
+ std::string makeParallelFileName(int id, int t);
+
+protected:
const std::string probeName;
const std::string outputPath;
@@ -202,7 +208,6 @@ private:
bool hasDeviceQuantityArray; //!> flag initiating memCopy in Point and PlaneProbe. Other probes are only based on thrust reduce functions and therefore dont need explict memCopy in interact()
bool outputTimeSeries; //!> flag initiating overwrite of output vtk files, skipping collection files and limiting the length of the written data to the current time step (currently only used for WallModelProbe)
std::vector<std::string> fileNamesForCollectionFile;
- std::vector<std::string> varNames;
bool fileNameLU = true; //!> if true, written file name contains time step in LU, else is the number of the written probe files
@@ -215,7 +220,6 @@ protected:
uint tProbe = 0; //!> counter for number of probe evaluations. Only used when outputting timeseries
-
std::function<real(int)> velocityRatio;
std::function<real(int)> densityRatio;
std::function<real(int)> forceRatio;
diff --git a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/WallModelProbe.h b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/WallModelProbe.h
index d6464c5ca2aa60310cc6bb7ca0a210bc12e755ff..4ea90f74c7a0d57af4995e1b5874234967f1e901 100644
--- a/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/WallModelProbe.h
+++ b/src/gpu/VirtualFluids_GPU/PreCollisionInteractor/Probes/WallModelProbe.h
@@ -55,14 +55,17 @@ public:
uint _tStartOut,
uint _tOut
): Probe(_probeName,
- _outputPath,
- _tStartAvg,
- _tStartTmpAvg,
- _tAvg,
- _tStartOut,
- _tOut,
- false,
- true){}
+ _outputPath,
+ _tStartAvg,
+ _tStartTmpAvg,
+ _tAvg,
+ _tStartOut,
+ _tOut,
+ false,
+ true)
+ {
+ if (_tStartTmpAvg<_tStartAvg) throw std::runtime_error("Probe: tStartTmpAvg must be larger than tStartAvg!");
+ }
void setForceOutputToStress(bool _outputStress){ this->outputStress = _outputStress; }
diff --git a/src/lbm/constants/NumericConstants.h b/src/lbm/constants/NumericConstants.h
index 4918d49aaa0431de639ea8ba3320c4fa45e539d4..b61e994e726e0b5cd85d8130bd6548ccebdc27f6 100644
--- a/src/lbm/constants/NumericConstants.h
+++ b/src/lbm/constants/NumericConstants.h
@@ -18,6 +18,7 @@ static constexpr double c1o8 = 0.125;
static constexpr double c1o9 = 0.111111111111111;
static constexpr double c2o9 = 0.222222222222222;
static constexpr double c4o9 = 0.444444444444444;
+static constexpr double c4o10 = 0.4;
static constexpr double c1o10 = 0.1;
static constexpr double c1o12 = 0.083333333333333;
static constexpr double c1o16 = 0.0625;
@@ -48,6 +49,7 @@ static constexpr double c99o100 = 0.99;
static constexpr double c1o126 = 0.007936507936508;
static constexpr double c1o216 = 0.004629629629630;
static constexpr double c5o4 = 1.25;
+static constexpr double c4o3 = 1.333333333333333;
static constexpr double c9o4 = 2.25;
static constexpr double c5o2 = 2.5;
static constexpr double c9o2 = 4.5;
@@ -99,15 +101,15 @@ static constexpr double c72o1 = 72.;
static constexpr double c84o1 = 84.;
static constexpr double c88o1 = 88.;
static constexpr double c96o1 = 96.;
-static constexpr double c100o1 = 10.;
-static constexpr double c130o1 = 13.;
-static constexpr double c152o1 = 15.;
-static constexpr double c166o1 = 16.;
-static constexpr double c195o1 = 19.;
-static constexpr double c216o1 = 21.;
-static constexpr double c264o1 = 26.;
-static constexpr double c290o1 = 29.;
-static constexpr double c367o1 = 36.;
+static constexpr double c100o1 = 100.;
+static constexpr double c130o1 = 130.;
+static constexpr double c152o1 = 152.;
+static constexpr double c166o1 = 166.;
+static constexpr double c195o1 = 195.;
+static constexpr double c216o1 = 216.;
+static constexpr double c264o1 = 264.;
+static constexpr double c290o1 = 290.;
+static constexpr double c367o1 = 367.;
static constexpr double Op0000002 = 0.0000002;
static constexpr double c10eM30 = 1e-30;
@@ -132,6 +134,7 @@ static constexpr float c1o8 = 0.125f;
static constexpr float c1o9 = (1.0f / 9.0f);
static constexpr float c2o9 = (2.0f / 9.0f);
static constexpr float c4o9 = (4.0f / 9.0f);
+static constexpr float c4o10 = 0.4f;
static constexpr float c1o10 = 0.1f;
static constexpr float c1o12 = (1.0f / 12.0f);
static constexpr float c1o16 = 0.0625f;
@@ -162,6 +165,7 @@ static constexpr float c99o100 = 0.99f;
static constexpr float c1o126 = (1.0f / 126.0f);
static constexpr float c1o216 = (1.0f / 216.0f);
static constexpr float c5o4 = 1.25f;
+static constexpr float c4o3 = (4.0f / 3.0f);
static constexpr float c9o4 = 2.25f;
static constexpr float c5o2 = 2.5f;
static constexpr float c9o2 = 4.5f;
diff --git a/utilities/setup_builder.py b/utilities/setup_builder.py
new file mode 100644
index 0000000000000000000000000000000000000000..94e4f805b3e05024742ead72e3ffb28c477c282f
--- /dev/null
+++ b/utilities/setup_builder.py
@@ -0,0 +1,34 @@
+from setuptools import build_meta
+
+class builder(build_meta._BuildMetaBackend):
+
+ def run_setup(self, setup_script='setup.py'):
+ # Note that we can reuse our build directory between calls
+ # Correctness comes first, then optimization later
+ __file__ = setup_script
+ __name__ = '__main__'
+
+ with build_meta._open_setup_script(__file__) as f:
+ code = f.read().replace(r'\r\n', r'\n')
+ args = locals()
+ args["config_args"] = self.extra_args
+ exec(code, args)
+
+
+ def add_settings(self, config_settings):
+ self.extra_args = dict()
+ if config_settings:
+ self.extra_args = config_settings
+
+ def build_wheel(self, wheel_directory, config_settings=None,
+ metadata_directory=None):
+ self.add_settings(config_settings)
+ return super().build_wheel(wheel_directory, config_settings, metadata_directory)
+
+ def build_sdist(self, sdist_directory, config_settings=None):
+ self.add_settings(config_settings)
+ return super().build_wheel(sdist_directory, config_settings)
+
+build = builder()
+build_wheel = build.build_wheel
+build_sdist = build.build_sdist
\ No newline at end of file