projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 5865476) | patch
[OpenMP][AMDGPU] Add 'uniform-work-group' attribute to OpenMP kernels
authorJoseph Huber <jhuber6@vols.utk.edu>
Thu, 6 Oct 2022 16:17:39 +0000 (11:17 -0500)
committerJoseph Huber <jhuber6@vols.utk.edu>
Thu, 6 Oct 2022 23:22:09 +0000 (18:22 -0500)
commit4aa87a131f93de73b8e245021af0294545d063c5
treed04852dbd4305a70d9db797506f97b9ff44919ddtree | snapshot
parent586547687946b27a99d3bb4241226f9f126a173ecommit | diff
[OpenMP][AMDGPU] Add 'uniform-work-group' attribute to OpenMP kernels

The `cl-uniform-work-group` attribute asserts that the global work-size
be a multiple of the work-group specified work group size. This should
allow optimizations. It is already present by default in the AMD
compiler and for HIP kernels so it should be safe to allow this for
OpenMP kernels by default.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D135374
clang/lib/CodeGen/TargetInfo.cpp diff | blob | history
clang/test/OpenMP/amdgcn-attributes.cpp diff | blob | history
Domain: System / Toolchain;