[all-commits] [llvm/llvm-project] 3df36a: [AMDGPU] Enable vectorization of i8 values. (#134934)

[Gheorghe-Teodor Bercea via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 3df36a2b1872ad49709fcfdfec69da57707f9286
      https://github.com/llvm/llvm-project/commit/3df36a2b1872ad49709fcfdfec69da57707f9286
  Author: Gheorghe-Teodor Bercea <doru.bercea at amd.com>
  Date:   2025-06-26 (Thu, 26 Jun 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
    M llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
    M llvm/test/Analysis/CostModel/AMDGPU/load.ll
    M llvm/test/Analysis/CostModel/AMDGPU/shufflevector.ll
    M llvm/test/Analysis/CostModel/AMDGPU/store.ll
    M llvm/test/Transforms/SLPVectorizer/AMDGPU/vectorize-i8.ll

  Log Message:
  -----------
  [AMDGPU] Enable vectorization of i8 values. (#134934)

This patch adjusts the cost model to account for the ability of the
AMDGPU optimizer to group together i8 values into i32 values.

Co-authored-by: Erich Keane <ekeane at nvidia.com>



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