[all-commits] [llvm/llvm-project] 369d78: [PowerPC] Optimal sequence for doubleword vec_all_...

[Nemanja Ivanovic via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 369d785574f5a22c086d0c40268a39a64bdd7217
      https://github.com/llvm/llvm-project/commit/369d785574f5a22c086d0c40268a39a64bdd7217
  Author: Nemanja Ivanovic <nemanja.i.ibm at gmail.com>
  Date:   2021-10-01 (Fri, 01 Oct 2021)

  Changed paths:
    M clang/lib/Headers/altivec.h
    M clang/test/CodeGen/builtins-ppc-vsx.c

  Log Message:
  -----------
  [PowerPC] Optimal sequence for doubleword vec_all_{eq|ne} on Power7

These builtins produce inefficient code for CPU's prior to Power8
due to vcmpequd being unavailable. The predicate forms can actually
leverage the available vcmpequw along with xxlxor to produce a better
sequence.