[all-commits] [llvm/llvm-project] f0e8c1: [CostModel][X86] Adjust 256-bit select costs to ac...

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: f0e8c1d6d99ef686fd381c6af4debc7470876681
      https://github.com/llvm/llvm-project/commit/f0e8c1d6d99ef686fd381c6af4debc7470876681
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2022-05-06 (Fri, 06 May 2022)

  Changed paths:
    M llvm/lib/Target/X86/X86TargetTransformInfo.cpp
    M llvm/test/Analysis/CostModel/X86/arith-sminmax.ll
    M llvm/test/Analysis/CostModel/X86/arith-ssat.ll
    M llvm/test/Analysis/CostModel/X86/arith-uminmax.ll
    M llvm/test/Analysis/CostModel/X86/arith-usat.ll
    M llvm/test/Analysis/CostModel/X86/fptoi_sat.ll
    M llvm/test/Analysis/CostModel/X86/fshl.ll
    M llvm/test/Analysis/CostModel/X86/fshr.ll
    M llvm/test/Analysis/CostModel/X86/reduce-smax.ll
    M llvm/test/Analysis/CostModel/X86/reduce-smin.ll
    M llvm/test/Analysis/CostModel/X86/reduce-umax.ll
    M llvm/test/Analysis/CostModel/X86/reduce-umin.ll
    M llvm/test/Analysis/CostModel/X86/vselect-cost.ll
    M llvm/test/Transforms/SLPVectorizer/X86/arith-add-ssat.ll
    M llvm/test/Transforms/SLPVectorizer/X86/arith-sub-ssat.ll

  Log Message:
  -----------
  [CostModel][X86] Adjust 256-bit select costs to account for slow BLENDV op

Based off the script from D103695, on AVX1, Jaguar/Bulldozer both have low throughput for ymm select patterns (BLENDV + OR(AND,ANDN))), and even on AVX2 Haswell still struggles with BLENDV ops