[all-commits] [llvm/llvm-project] c24d68: [AMDGPU] Take advantage of VOP3 literals in conver...

[Jay Foad via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: c24d68fff1fe8d3115c411d6e81092eb1f855b52
      https://github.com/llvm/llvm-project/commit/c24d68fff1fe8d3115c411d6e81092eb1f855b52
  Author: Jay Foad <jay.foad at amd.com>
  Date:   2022-08-02 (Tue, 02 Aug 2022)

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

  Log Message:
  -----------
  [AMDGPU] Take advantage of VOP3 literals in convertToThreeAddress

This improves a corner case where v_fmac can be converted to v_fma on
GFX10+ even if it has a literal operand.

Differential Revision: https://reviews.llvm.org/D130992