[all-commits] [llvm/llvm-project] 004bf1: [AArch64] Emit FNMADD instead of FNEG(FMADD)

[Matthew Devereau via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 004bf170c6cbaa049601bcf92f86a9459aec2dc2
      https://github.com/llvm/llvm-project/commit/004bf170c6cbaa049601bcf92f86a9459aec2dc2
  Author: Matt Devereau <matthew.devereau at arm.com>
  Date:   2023-05-10 (Wed, 10 May 2023)

  Changed paths:
    M llvm/include/llvm/CodeGen/MachineCombinerPattern.h
    M llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
    A llvm/test/CodeGen/AArch64/aarch64_fnmadd.ll

  Log Message:
  -----------
  [AArch64] Emit FNMADD instead of FNEG(FMADD)

Emit FNMADD instead of FNEG(FMADD) for optimization levels
above Oz when fast-math flags (nsz+contract) permit it.

Differential Revision: https://reviews.llvm.org/D149260