[all-commits] [llvm/llvm-project] fe6c11: [CostModel][X86] Improve f64/v2f64/v4f64 FMUL cost...

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: fe6c11c5710e95eedbb16c2aac58a5d992e55434
      https://github.com/llvm/llvm-project/commit/fe6c11c5710e95eedbb16c2aac58a5d992e55434
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2021-05-21 (Fri, 21 May 2021)

  Changed paths:
    M llvm/lib/Target/X86/X86TargetTransformInfo.cpp
    M llvm/test/Analysis/CostModel/X86/arith-fp.ll
    M llvm/test/Analysis/CostModel/X86/reduce-fmul.ll
    M llvm/test/Transforms/SLPVectorizer/X86/simplebb.ll

  Log Message:
  -----------
  [CostModel][X86] Improve f64/v2f64/v4f64 FMUL costs on AVX1 targets to account for slower btver2

BTVER2 has a weaker f64 multiplier that other AVX1-era targets, so we need to bump the worst case cost slightly - llvm-mca reports the new vectorization in simplebb is beneficial on btver2, bdver2 and sandybridge AVX1 targets