[PATCH] D152191: AMDGPU: Fix regressions in integer mad matching

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: AMDGPU, cfang, Pierre-vh, rampitec, foad.
Herald added subscribers: StephenFan, kerbowa, hiraditya, tpr, dstuttard, yaxunl, jvesely, kzhuravl.
Herald added a project: All.
arsenm requested review of this revision.
Herald added a subscriber: wdng.
Herald added a project: LLVM.

Undo the canonicalize done in
0cfc6510323fbb5a56a5de23cbc65f7cc30fd34c <https://reviews.llvm.org/rG0cfc6510323fbb5a56a5de23cbc65f7cc30fd34c>. Restores some regressed
matching of integer mad. The selection patterns fo the actual mads
don't seem to be properly commuting, so some of the commuted cases are
still missed.

      

Fixes: SWDEV-363009


https://reviews.llvm.org/D152191

Files:
  llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
  llvm/test/CodeGen/AMDGPU/reassoc-mul-add-1-to-mad.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D152191.528555.patch
Type: text/x-patch
Size: 54379 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230605/188ddb7a/attachment.bin>