automatically set para->useStreams to false for single gpu

In copyEdgeNodes the nodes were copied to the wrong places of the send vector. The problem was that the full vector sendProcessNeighborHostAllNodes was used for copying instead of the shorter vector sendProcessNeighborHostAfterFtoC

Only copy send nodes which are not part of the reduced communication (for reduced comm after fine to coarse)

for debugging

used https://kitware.github.io/vtk-examples/site/VTKFileFormats/ as guide

Anna Wellmann authored 3 years ago and Anna Wellmann committed 3 years ago

The scaling coarse to fine for bulk nodes should not overlap with the collision kernel for boder nodes on the fine level .

(cherry picked from commit dafce4df)

The scaling coarse to fine for bulk nodes should not overlap with the collision kernel for boder nodes on the fine level .

Anna Wellmann authored 3 years ago and Anna Wellmann committed 3 years ago

(cherry picked from commit fc68e403)