[all-commits] [llvm/llvm-project] e3c932: [X86][CodeGen] Set isReMaterializable = 1 for AVX ...

[Shengchen Kan via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: e3c9327bc493286bf420d1520df8217ae559f5c3
      https://github.com/llvm/llvm-project/commit/e3c9327bc493286bf420d1520df8217ae559f5c3
  Author: Shengchen Kan <shengchen.kan at intel.com>
  Date:   2024-01-31 (Wed, 31 Jan 2024)

  Changed paths:
    M llvm/lib/Target/X86/X86InstrInfo.cpp
    M llvm/lib/Target/X86/X86InstrSSE.td
    M llvm/test/CodeGen/X86/matrix-multiply.ll
    M llvm/test/CodeGen/X86/vector-interleaved-store-i8-stride-8.ll

  Log Message:
  -----------
  [X86][CodeGen] Set isReMaterializable = 1 for AVX broadcast load

Broadcast of a single float should not be any slower than
loading 32B using vmovaps. So remat it can help reduce
register spill when there is big register pressure.