[all-commits] [llvm/llvm-project] ca4098: AMDGPU: Permit more frame index operands in verifi...

[Matt Arsenault via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: ca409892c5396fa3fbb8ea4dbf53d0e952f36d09
      https://github.com/llvm/llvm-project/commit/ca409892c5396fa3fbb8ea4dbf53d0e952f36d09
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-08-05 (Mon, 05 Aug 2024)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
    A llvm/test/CodeGen/AMDGPU/eliminate-frame-index-s-add-i32.mir
    A llvm/test/CodeGen/AMDGPU/eliminate-frame-index-scalar-bit-ops.mir
    A llvm/test/CodeGen/AMDGPU/eliminate-frame-index-v-add-co-u32.mir
    A llvm/test/CodeGen/AMDGPU/eliminate-frame-index-v-add-u32.mir

  Log Message:
  -----------
  AMDGPU: Permit more frame index operands in verifier (#101691)

Treat FI operands more like a register. When it gets materialized,
we will typically need to introduce a scavenged register anyway.
Add baseline tests for folding frame indexes into add/or.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications