Shuenhoy/vtk-m
1
0
Fork 0
mirror of https://gitlab.kitware.com/vtk/vtk-m synced 2024-09-19 18:45:43 +00:00
Code Issues 2 Actions Packages Projects Releases Wiki Activity

Merge topic 'better_3dschedule_for_summit'

Browse Source 
6cdf6cb67 Less aggressive defaults for VTK-m compared to summit.
387947918 Improve VTK-m cuda scheduling based on Summit scaling study

Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Kenneth Moreland <kmorel@sandia.gov>
Merge-request: !1576
This commit is contained in:
Robert Maynard 2019-03-20 20:53:50 +00:00 committed by Kitware Robot
commit a7beb532ea
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
vtkm/cont/cuda/internal/DeviceAdapterAlgorithmCuda.cu
View File

@ -125,18 +125,18 @@ void DeviceAdapterAlgorithm<vtkm::cont::DeviceAdapterTagCuda>::GetGridsAndBlocks
if (size.x == 0)
{ //grids that have no x dimension
blocks.x = 1;
blocks.y = 16;
blocks.y = 8;
blocks.z = 8;
}
else if (size.x > 128)
{
blocks.x = 64;
blocks.y = 2;
blocks.z = 1;
blocks.x = 8;
blocks.y = 8;
blocks.z = 4;
}
else
{ //for really small grids
blocks.x = 8;
blocks.x = 4;
blocks.y = 4;
blocks.z = 4;
}