[all-commits] [llvm/llvm-project] 32073b: AMDGPU: Do not generate non-temporal hint when Loa...

[Changpeng Fang via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 32073b835674a9e7bc3e1ee9708efb7c58e7394f
      https://github.com/llvm/llvm-project/commit/32073b835674a9e7bc3e1ee9708efb7c58e7394f
  Author: Changpeng Fang <changpeng.fang at amd.com>
  Date:   2024-01-23 (Tue, 23 Jan 2024)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIISelLowering.cpp
    M llvm/test/CodeGen/AMDGPU/llvm.amdgcn.global.load.tr-w32.ll
    M llvm/test/CodeGen/AMDGPU/llvm.amdgcn.global.load.tr-w64.ll

  Log Message:
  -----------
  AMDGPU: Do not generate non-temporal hint when Load_Tr intrinsic did not specify it (#79104)

int_amdgcn_global_load_tr did not specify non-temporal load transpose,
thus we should
not genetrate the non-temporal hint for the load. We need to implement
getTgtMemIntrinsic
to create the corresponding MemSDNode. And we don't set the non-temporal
flag because
the intrinsic did not specify it.

NOTE: We need to implement getTgtMemIntrinsic for any memory intrinsics.