[PATCH] D112827: [AMDGPU][GlobalISel] Fold G_FNEG above when users cannot fold mods

[Mirko Brkusanin via Phabricator via llvm-commits]

mbrkusanin created this revision.
mbrkusanin added reviewers: foad, arsenm.
mbrkusanin added a project: LLVM.
Herald added subscribers: kerbowa, hiraditya, t-tye, tpr, dstuttard, rovka, yaxunl, nhaehnle, jvesely, kzhuravl.
mbrkusanin requested review of this revision.
Herald added a subscriber: wdng.

If possible fold fneg into instruction above if users cannot fold mods and we
know it will decrease instruction count.
Follows same logic as SDAG combiner in choosing opportunities to combine.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D112827

Files:
  llvm/include/llvm/CodeGen/GlobalISel/CombinerHelper.h
  llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp
  llvm/lib/Target/AMDGPU/AMDGPUCombine.td
  llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp
  llvm/test/CodeGen/AMDGPU/GlobalISel/combine-foldable-fneg.mir

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D112827.383383.patch
Type: text/x-patch
Size: 42466 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20211029/929192d3/attachment.bin>