[all-commits] [llvm/llvm-project] 92e0ab: [AArch64] Don't map llvm sqrt intrinsics to veclib...

[huntergr-arm via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 92e0ab937ffd348e723586ac0a980ffeca4dad48
      https://github.com/llvm/llvm-project/commit/92e0ab937ffd348e723586ac0a980ffeca4dad48
  Author: Graham Hunter <graham.hunter at arm.com>
  Date:   2023-03-07 (Tue, 07 Mar 2023)

  Changed paths:
    M llvm/include/llvm/Analysis/VecFuncs.def
    A llvm/test/CodeGen/AArch64/replace-intrinsics-with-veclib-sqrt.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/sleef-calls-aarch64.ll

  Log Message:
  -----------
  [AArch64] Don't map llvm sqrt intrinsics to veclib functions

Since AArch64 has sqrt instructions, we want to use those instead of
calls to vector math routines for llvm sqrt intrinsics (since those
don't imply some of the constraints that libm calls might have) so
we just remove the mappings.

Code originally written by mgabka

Reviewed By: danielkiss, paulwalker-arm

Differential Revision: https://reviews.llvm.org/D145392