[PATCH] D110557: [X86] Optimize fdiv with reciprocal instructions for half type
Simon Pilgrim via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Oct 5 08:46:49 PDT 2021
RKSimon added inline comments.
================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:23187
+ SDValue Undef = DAG.getUNDEF(MVT::v8f16);
+ Op = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, MVT::v8f16, Undef, Op, Zero);
+ Op = DAG.getNode(X86ISD::RSQRT14S, DL, MVT::v8f16, Undef, Op);
----------------
This is a SCALAR_TO_VECTOR node
================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:23236
+ SDValue Undef = DAG.getUNDEF(MVT::v8f16);
+ Op = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, MVT::v8f16, Undef, Op, Zero);
+ Op = DAG.getNode(X86ISD::RCP14S, DL, MVT::v8f16, Undef, Op);
----------------
SCALAR_TO_VECTOR
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D110557/new/
https://reviews.llvm.org/D110557
More information about the llvm-commits
mailing list