[all-commits] [llvm/llvm-project] 380087: [AMDGPU] Add test with redundant copies to tempora...

[Mirko Brkušanin via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: 380087e6c9a226b28c24de3b18e202e20d430765
      https://github.com/llvm/llvm-project/commit/380087e6c9a226b28c24de3b18e202e20d430765
  Author: Mirko Brkusanin <Mirko.Brkusanin at amd.com>
  Date:   2020-10-08 (Thu, 08 Oct 2020)

  Changed paths:
    A llvm/test/CodeGen/AMDGPU/load-local-redundant-copies.ll

  Log Message:
  -----------
  [AMDGPU] Add test with redundant copies to temporary stack slot produced by expandUnalignedLoad

Differential Revision: https://reviews.llvm.org/D88895


  Commit: 7c88d13fd1c1a864124d96224996c62192d431f2
      https://github.com/llvm/llvm-project/commit/7c88d13fd1c1a864124d96224996c62192d431f2
  Author: Mirko Brkusanin <Mirko.Brkusanin at amd.com>
  Date:   2020-10-08 (Thu, 08 Oct 2020)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIISelLowering.cpp
    M llvm/test/CodeGen/AMDGPU/load-local-redundant-copies.ll

  Log Message:
  -----------
  [AMDGPU] Prefer SplitVectorLoad/Store over expandUnalignedLoad/Store

ExpandUnalignedLoad/Store can sometimes produce unnecessary copies to
temporary stack slot. We should prefer splitting vectors if possible.

Differential Revision: https://reviews.llvm.org/D88882


Compare: https://github.com/llvm/llvm-project/compare/a96bcfb19674...7c88d13fd1c1