[all-commits] [llvm/llvm-project] 9f4746: AMDGPU: Combine down fcopysign f64 magnitude

[Matt Arsenault via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 9f4746b65f9fdea39975abeea37fd1f96475d369
      https://github.com/llvm/llvm-project/commit/9f4746b65f9fdea39975abeea37fd1f96475d369
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2023-03-06 (Mon, 06 Mar 2023)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIISelLowering.cpp
    M llvm/test/CodeGen/AMDGPU/fcopysign.f16.ll
    M llvm/test/CodeGen/AMDGPU/fcopysign.f64.ll
    M llvm/test/CodeGen/AMDGPU/llvm.round.f64.ll

  Log Message:
  -----------
  AMDGPU: Combine down fcopysign f64 magnitude

Copy through the low bits and only apply an f32
copysign to the high half. This is effectively
what we do for codegen anyway, but this provides
some combine benefits. The cases involving constants
show some small improvements.

https://reviews.llvm.org/D142682