[all-commits] [llvm/llvm-project] 05e7d8: TTI: Add addrspace parameters to memcpy lowering f...

[Matt Arsenault via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: 05e7d8d6ce0e4f87aa1c4490d9f843360a6b2ca5
      https://github.com/llvm/llvm-project/commit/05e7d8d6ce0e4f87aa1c4490d9f843360a6b2ca5
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2020-03-16 (Mon, 16 Mar 2020)

  Changed paths:
    M llvm/include/llvm/Analysis/TargetTransformInfo.h
    M llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
    M llvm/lib/Analysis/TargetTransformInfo.cpp
    M llvm/lib/Transforms/Utils/LowerMemIntrinsics.cpp

  Log Message:
  -----------
  TTI: Add addrspace parameters to memcpy lowering functions


  Commit: b0bdb186f506d7fa2686f1c09bb8926b6adeb91d
      https://github.com/llvm/llvm-project/commit/b0bdb186f506d7fa2686f1c09bb8926b6adeb91d
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2020-03-16 (Mon, 16 Mar 2020)

  Changed paths:
    M llvm/lib/Transforms/Utils/LowerMemIntrinsics.cpp
    M llvm/test/CodeGen/AMDGPU/lower-mem-intrinsics.ll

  Log Message:
  -----------
  Utils: Always set alignment when expanding mem intrinsics

This was creating natural aligned loads and stores, which may not be
the case. The target could request a wider type load with less
alignment.


Compare: https://github.com/llvm/llvm-project/compare/7d382dcd46a1...b0bdb186f506