[PATCH] D106891: [AMDGPU] [Remarks] Emit optimization remarks when an FP atomic instruction is converted into a CAS loop or unsafe hardware instruction for GFX90A

Anshil Gandhi via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Tue Jul 27 10:47:49 PDT 2021


gandhi21299 updated this revision to Diff 362093.
gandhi21299 added a comment.

- replaced `check-prefixes` to `check-prefix` in the test
- Function ffp1 emits a `flat_atomic_cmpswap` instead of `global_atomic_add`, I am not sure if that is considered an unsafe hardware instruction. @arsenm ?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D106891/new/

https://reviews.llvm.org/D106891

Files:
  clang/test/CodeGenCUDA/fp-atomics-optremarks.cu
  llvm/lib/Target/AMDGPU/SIISelLowering.cpp
  llvm/lib/Target/AMDGPU/SIISelLowering.h

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D106891.362093.patch
Type: text/x-patch
Size: 6880 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20210727/b26cb46f/attachment-0001.bin>


More information about the cfe-commits mailing list