[all-commits] [llvm/llvm-project] 9962eb: [CostModel][X86] Add vXi32 multiply pattern tests

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 9962ebaee5822556a8b5105796f258979c56af81
      https://github.com/llvm/llvm-project/commit/9962ebaee5822556a8b5105796f258979c56af81
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2021-09-05 (Sun, 05 Sep 2021)

  Changed paths:
    A llvm/test/Analysis/CostModel/X86/mul32.ll

  Log Message:
  -----------
  [CostModel][X86] Add vXi32 multiply pattern tests

Add tests for vXi32 multiplies where the operands have been extended from vXi8/vXi16


  Commit: f114ef3731dfd79e8f235cc02ec9879c489ebf96
      https://github.com/llvm/llvm-project/commit/f114ef3731dfd79e8f235cc02ec9879c489ebf96
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2021-09-05 (Sun, 05 Sep 2021)

  Changed paths:
    M llvm/lib/Target/X86/X86TargetTransformInfo.cpp
    M llvm/test/Analysis/CostModel/X86/mul32.ll
    M llvm/test/Analysis/CostModel/X86/slm-arith-costs.ll
    M llvm/test/Transforms/LoopVectorize/X86/mul_slm_16bit.ll

  Log Message:
  -----------
  [CostModel][X86] Add generic costs for vXi32 MUL -> v2Xi16 PMADDDW folds

Based off the improved fold in D108522

This should eventually allow us to replace the SLM only cost patterns with generic versions.


Compare: https://github.com/llvm/llvm-project/compare/8523fb96a63e...f114ef3731df