[all-commits] [llvm/llvm-project] 95cb4d: AMDGPU/GlobalISel: RegBankLegalize rules for wave_...

[vangthao95 via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 95cb4d4b536af5c6f462e7a2c6ad2dd14c2701b7
      https://github.com/llvm/llvm-project/commit/95cb4d4b536af5c6f462e7a2c6ad2dd14c2701b7
  Author: vangthao95 <vang.thao at amd.com>
  Date:   2026-05-26 (Tue, 26 May 2026)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp
    M llvm/test/CodeGen/AMDGPU/llvm.amdgcn.wave.shuffle.ll

  Log Message:
  -----------
  AMDGPU/GlobalISel: RegBankLegalize rules for wave_shuffle (#196412)

Add uniform and divergent RegBankLegalize rules for wave shuffle. When
wave_shuffle is uniform, it is folded by AMDGPUUniformIntrinsicCombine
before reaching RegBankLegalize, to its value operand when value is
uniform, or to amdgcn_readlane when only idx is uniform. Add a test to
disable UniformIntrinsicCombine so we can verify and test for uniform
rule.



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