[PATCH] D27419: [X86][SSE] Consistently set MOVD/MOVQ load/store/move instructions to integer domain

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

Should the ones that use FR32 and FR64 in their types stay in floating point domain? Doesn't that make them likely to be near floating point code?


Repository:
  rL LLVM

https://reviews.llvm.org/D27419