[all-commits] [llvm/llvm-project] 31bd3e: [AMDGPU] Use VGPR classes in divergent build_vecto...

[Jay Foad via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 31bd3e6e92e6249c2741219b24830abf2213bd22
      https://github.com/llvm/llvm-project/commit/31bd3e6e92e6249c2741219b24830abf2213bd22
  Author: Jay Foad <jay.foad at amd.com>
  Date:   2022-10-21 (Fri, 21 Oct 2022)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInstructions.td

  Log Message:
  -----------
  [AMDGPU] Use VGPR classes in divergent build_vector patterns

This does not appear to affect codegen, but using SGPR classes for
operands of VALU instructions looked weird.

Differential Revision: https://reviews.llvm.org/D136459