[all-commits] [llvm/llvm-project] 9ac3a8: [AMDGPU] Disentangle MFMA handling in convertToThr...

[Jay Foad via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 9ac3a8504736dc1dbe249a31ebbaa0f861e576eb
      https://github.com/llvm/llvm-project/commit/9ac3a8504736dc1dbe249a31ebbaa0f861e576eb
  Author: Jay Foad <jay.foad at amd.com>
  Date:   2022-03-01 (Tue, 01 Mar 2022)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInstrInfo.cpp

  Log Message:
  -----------
  [AMDGPU] Disentangle MFMA handling in convertToThreeAddress. NFC.

Move MFMA handling to the top of convertToThreeAddress and pull
IsF16 calculation out of the switch. I think this makes it clearer
exactly which mac/fmac opcodes are handled, since they are now
listed in the switch with minimal extra clutter.

Differential Revision: https://reviews.llvm.org/D120703


  Commit: 289339140e7121d6ed48a3bd556eb429163e04b0
      https://github.com/llvm/llvm-project/commit/289339140e7121d6ed48a3bd556eb429163e04b0
  Author: Jay Foad <jay.foad at amd.com>
  Date:   2022-03-01 (Tue, 01 Mar 2022)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
    M llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.fmul.legacy.ll
    M llvm/test/CodeGen/AMDGPU/llvm.amdgcn.fma.legacy.ll
    M llvm/test/CodeGen/AMDGPU/llvm.amdgcn.fmul.legacy.ll

  Log Message:
  -----------
  [AMDGPU] Handle legacy multiply-accumulate opcodes in convertToThreeAddress

Handle V_MAC_LEGACY_F32 and V_FMAC_LEGACY_F32 in
convertToThreeAddress, to avoid the need for an extra mov
instruction in some cases.

Differential Revision: https://reviews.llvm.org/D120704


Compare: https://github.com/llvm/llvm-project/compare/9ffeaaa0ea54...289339140e71