[all-commits] [llvm/llvm-project] 0176ce: AMDGPU: Fold bitcasts into readfirstlane, readlane...

[Matt Arsenault via All-commits]

  Branch: refs/heads/users/arsenm/amdgpu/fold-bitcasts-readfirstlane
  Home:   https://github.com/llvm/llvm-project
  Commit: 0176ce4702134bcc1407f61393c3ba5da7a4b2c6
      https://github.com/llvm/llvm-project/commit/0176ce4702134bcc1407f61393c3ba5da7a4b2c6
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2025-02-26 (Wed, 26 Feb 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
    M llvm/test/Transforms/InstCombine/AMDGPU/bitcast-fold-lane-ops.ll
    M llvm/test/Transforms/InstCombine/AMDGPU/permlane64.ll

  Log Message:
  -----------
  AMDGPU: Fold bitcasts into readfirstlane, readlane, and permlane64

We should handle this for all the handled readlane and dpp ops.


  Commit: 7bd0159075b8509e77dc720d7ce75ebe50c7d6f6
      https://github.com/llvm/llvm-project/commit/7bd0159075b8509e77dc720d7ce75ebe50c7d6f6
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2025-02-26 (Wed, 26 Feb 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp

  Log Message:
  -----------
  Make sure convergence tokens are preserved


  Commit: db5757a8b36d981455a54880bdc8a70f6629822d
      https://github.com/llvm/llvm-project/commit/db5757a8b36d981455a54880bdc8a70f6629822d
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2025-02-26 (Wed, 26 Feb 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp

  Log Message:
  -----------
  Revert "Make sure convergence tokens are preserved"

This reverts commit bf1798782624bd60b22d7d866773e0cd5b5b1c75.


  Commit: 2b4c5903614ee3f18d80ae084a7ccb36234574f4
      https://github.com/llvm/llvm-project/commit/2b4c5903614ee3f18d80ae084a7ccb36234574f4
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2025-02-26 (Wed, 26 Feb 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp

  Log Message:
  -----------
  Use bundle guard


  Commit: ad6c4d9af31e3515a2c9730f851fb36729a16ac9
      https://github.com/llvm/llvm-project/commit/ad6c4d9af31e3515a2c9730f851fb36729a16ac9
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2025-02-26 (Wed, 26 Feb 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
    M llvm/test/Transforms/InstCombine/AMDGPU/bitcast-fold-lane-ops.ll
    M llvm/test/Transforms/InstCombine/AMDGPU/permlane64.ll

  Log Message:
  -----------
  Using CallInst directly actually works


Compare: https://github.com/llvm/llvm-project/compare/a231960a9d8e...ad6c4d9af31e

To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications