[all-commits] [llvm/llvm-project] f246b1: AMDGPU: Add baseline tests for known sign bits of ...

[Matt Arsenault via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: f246b139bdbebe5f7f488cb38213b0a34608a147
      https://github.com/llvm/llvm-project/commit/f246b139bdbebe5f7f488cb38213b0a34608a147
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2022-12-05 (Mon, 05 Dec 2022)

  Changed paths:
    A llvm/test/CodeGen/AMDGPU/load-range-metadata-sign-bits.ll

  Log Message:
  -----------
  AMDGPU: Add baseline tests for known sign bits of load range metadata


  Commit: 47c68904a53a5aeeb2f7d506ac6be7d1a0fb951e
      https://github.com/llvm/llvm-project/commit/47c68904a53a5aeeb2f7d506ac6be7d1a0fb951e
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2022-12-05 (Mon, 05 Dec 2022)

  Changed paths:
    M llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
    M llvm/test/CodeGen/AMDGPU/load-range-metadata-sign-bits.ll

  Log Message:
  -----------
  DAG: ComputeNumSignBits from load range metadata

The cases where the result type doesn't match the range type
are inadequately tested, but I'm not sure how to write such a
test. During the pre-legalize combine, any obviously optimizable
code gets handled so it's harder to test legalized extloads.


Compare: https://github.com/llvm/llvm-project/compare/7969ab85e0a4...47c68904a53a