[PATCH] D64760: [x86] use more phadd for reductions
Simon Pilgrim via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Mon Jul 15 11:15:20 PDT 2019
RKSimon added inline comments.
================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:35618
+ if ((((VecVT == MVT::v8i16 || VecVT == MVT::v4i32) && Subtarget.hasSSSE3()) ||
+ ((VecVT == MVT::v4f32 || VecVT == MVT::v2f64) && Subtarget.hasSSE3()))) {
+ // TODO: Allow FADD with reduction and/or reassociation and no-signed-zeros.
----------------
(style) early-out if this is not the case
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D64760/new/
https://reviews.llvm.org/D64760
More information about the llvm-commits
mailing list