[PATCH] D88396: [X86] Replace movaps with movups when avx is enabled.

Craig Topper via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Wed Oct 7 08:07:57 PDT 2020


craig.topper added inline comments.


================
Comment at: llvm/lib/Target/X86/X86MovapsToMovups.cpp:1
+//===- X86MovapsToMovups.cpp ----------------------------------------------===//
+//
----------------
RKSimon wrote:
> Very minor issue - but this isn't just movups - how about X86UnalignedVectorMoves.cpp ?
But the patch only looks at movaps opcodes. It should look at all move opcodes.


================
Comment at: llvm/lib/Target/X86/X86MovapsToMovups.cpp:38
+static cl::opt<bool> EnableX86MovapsToMovups(
+    "enable-x86-movaps-to-movups", cl::Hidden,
+    cl::desc("X86: Enable transforming movaps instruction to movups."),
----------------
This option should start with x86-  and again it shouldn't just be movaps. It also needs to handle movapd movdqa movdqa64 and movdqa32.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D88396/new/

https://reviews.llvm.org/D88396



More information about the llvm-commits mailing list