[PATCH] D64981: [AMDGPU] Save some work when an atomic op has no uses

[Jay Foad via Phabricator via llvm-commits]

foad created this revision.
foad added reviewers: arsenm, dstuttard, tpr.
Herald added subscribers: jfb, hiraditya, t-tye, yaxunl, nhaehnle, wdng, jvesely, kzhuravl.
Herald added a project: LLVM.

In the atomic optimizer, save doing a bunch of work and generating a
bunch of dead IR in the fairly common case where the result of an
atomic op (i.e. the value that was in memory before the atomic op was
performed) is not used.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D64981

Files:
  llvm/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D64981.210776.patch
Type: text/x-patch
Size: 6401 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20190719/80572327/attachment.bin>