[all-commits] [llvm/llvm-project] 13fd73: [X86] znver2 - (V)EXTRACTPSrr takes 2 uops

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 13fd7373b6a95f014fa8ccb277be79656c248d14
      https://github.com/llvm/llvm-project/commit/13fd7373b6a95f014fa8ccb277be79656c248d14
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2022-11-20 (Sun, 20 Nov 2022)

  Changed paths:
    M llvm/lib/Target/X86/X86ScheduleZnver2.td
    M llvm/test/tools/llvm-mca/X86/Znver2/resources-avx1.s
    M llvm/test/tools/llvm-mca/X86/Znver2/resources-sse41.s

  Log Message:
  -----------
  [X86] znver2 - (V)EXTRACTPSrr takes 2 uops

D138359 was reporting that the EXTRACTPSrr override was unnecessary, however the AMD SoG and Agner both confirm that both the rr and rm versions take 2uops (matching znver1)


  Commit: 94d240a44a52dc4dfed52e0719ab4a728cdbdb4d
      https://github.com/llvm/llvm-project/commit/94d240a44a52dc4dfed52e0719ab4a728cdbdb4d
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2022-11-20 (Sun, 20 Nov 2022)

  Changed paths:
    M llvm/lib/Target/X86/X86SchedIceLake.td

  Log Message:
  -----------
  [X86] Remove unnecessary zmm shuffle instruction overrides from IceLake model

Reported by D138359 and confirmed with Intel AoM, Agner + uops.info


Compare: https://github.com/llvm/llvm-project/compare/f49c2a8c3659...94d240a44a52