[all-commits] [llvm/llvm-project] 06a4c8: Use v16i8 rather than v2i64 as the VT for memset e...

[Owen Anderson via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 06a4c858901d72389e01fbdd0f83b03c74d56831
      https://github.com/llvm/llvm-project/commit/06a4c858901d72389e01fbdd0f83b03c74d56831
  Author: Owen Anderson <resistor at mac.com>
  Date:   2021-08-19 (Thu, 19 Aug 2021)

  Changed paths:
    M llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
    A llvm/test/CodeGen/AArch64/memset.ll

  Log Message:
  -----------
  Use v16i8 rather than v2i64 as the VT for memset expansion on AArch64.

This allows the instruction selector to realize that it can directly
broadcast the low byte of the memset value, rather than replicating
it to a 64-bit GPR before broadcasting.

This fixes PR50985.

Differential Revision: https://reviews.llvm.org/D108354