opened
merge request
!27
"Fix clang-format off directive for older LLVM versions"
at
pycodegen / pystencils-sfg
accepted
merge request
!26
"Fixes to postprocessing: Remove unused code, test vector extraction, unify treatment of scalar fields"
at
pycodegen / pystencils-sfg
-
b1c47558 · Merge branch 'fhennig/postprocessing-fixes' into 'master'
- ... and 1 more commit. Compare 020347e7...b1c47558
opened
merge request
!26
"Fixes to postprocessing: Remove unused code, test vector extraction, unify treatment of scalar fields"
at
pycodegen / pystencils-sfg
accepted
merge request
!24
"Extend Support for CUDA and HIP kernel invocations"
at
pycodegen / pystencils-sfg
-
020347e7 · Merge branch 'fhennig/cuda-invoke' into 'master'
- ... and 1 more commit. Compare 8949bedb...020347e7
commented on
merge request !24
"Extend Support for CUDA and HIP kernel invocations"
at
pycodegen / pystencils-sfg
As far as testing is concerned, the CudaKernels and HipKernels test cases already run all available launch configurations at least once.
commented on
merge request !24
"Extend Support for CUDA and HIP kernel invocations"
at
pycodegen / pystencils-sfg
very good point. I pushed a patch wrapping the entire logic into a Builder class.
commented on
merge request !24
"Extend Support for CUDA and HIP kernel invocations"
at
pycodegen / pystencils-sfg
This gpu_invoke function is currently around 130 lines long and defines 3 local functions. From my perspective this is a bit complex and it took me...