Create helper class for creating subdomains on multiple GPUs
Previously it was very taxing to create apps that run on multiple GPUs. This merge aims to improve the situation.
Task before merge:
-
code review -
run regression tests for DrivenCavityMultiGPU (should pass) -
4 GPU -
8 GPU
-
-
test the SphereScaling app on the Phoenix cluster (I changed the app to a simpler setup) -
update regression tests for SphereScaling -
4 GPU -
8 GPU
-
-
standard pipelines pass
Edited by Anna Wellmann