[all-commits] [llvm/llvm-project] 9922d8: AMDGPU: Stop using aligned VGPR classes for addReg...

[Matt Arsenault via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 9922d8dd799dd22f3e34e774ed2750df56a4ca4d
      https://github.com/llvm/llvm-project/commit/9922d8dd799dd22f3e34e774ed2750df56a4ca4d
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2025-10-10 (Fri, 10 Oct 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIISelLowering.cpp
    M llvm/test/CodeGen/AMDGPU/mfma-loop.ll

  Log Message:
  -----------
  AMDGPU: Stop using aligned VGPR classes for addRegisterClass (#158278)

This is unnecessary. At use emission time, InstrEmitter will
use the common subclass of the value type's register class and
the use instruction register classes. This removes one of the
obstacles to treating special case instructions that do not have
the alignment requirement overly conservatively.



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