[PATCH] D122850: [AMDGPU] Fix regression with vectorization limiting

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

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

D67148 <https://reviews.llvm.org/D67148> has removed TTI::getNumberOfRegisters(bool Vector) and
started to call TTI::getNumberOfRegisters(unsigned ClassID) from
the LoopVectorize. This has resulted in an unrestricted vectorization
on AMDGPU blowing up register pressure.


https://reviews.llvm.org/D122850

Files:
  llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
  llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
  llvm/test/Transforms/LoopVectorize/AMDGPU/packed-fp32.ll
  llvm/test/Transforms/LoopVectorize/AMDGPU/packed-math.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D122850.419574.patch
Type: text/x-patch
Size: 17303 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220331/53051fd7/attachment.bin>