[all-commits] [llvm/llvm-project] ba3d6e: [AMDGPU] Rematerialize scalar loads (#68778)

[Piotr Sobczak via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: ba3d6e0499231c5f425b92cc126116b0e184dbf1
      https://github.com/llvm/llvm-project/commit/ba3d6e0499231c5f425b92cc126116b0e184dbf1
  Author: Piotr Sobczak <piotr.sobczak at amd.com>
  Date:   2023-10-26 (Thu, 26 Oct 2023)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
    M llvm/lib/Target/AMDGPU/SIInstrInfo.h
    M llvm/test/CodeGen/AMDGPU/hsa-metadata-kernel-code-props-v3.ll
    M llvm/test/CodeGen/AMDGPU/remat-smrd.mir
    M llvm/test/CodeGen/AMDGPU/snippet-copy-bundle-regression.mir

  Log Message:
  -----------
  [AMDGPU] Rematerialize scalar loads (#68778)

Extend the list of instructions that can be rematerialized in
SIInstrInfo::isReallyTriviallyReMaterializable() to support scalar
loads.

Try shrinking instructions to remat only the part needed for current
context. Add SIInstrInfo::reMaterialize target hook, and handle
shrinking of S_LOAD_DWORDX16_IMM to S_LOAD_DWORDX8_IMM as a proof of
concept.