[PATCH] D78986: [PowerPC-QPX] adjust operands order of qpx vector fma instructions.
Qing Shan Zhang via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Wed Apr 29 01:34:31 PDT 2020
steven.zhang added a subscriber: qiucf.
steven.zhang added inline comments.
================
Comment at: llvm/lib/Target/PowerPC/PPCInstrQPX.td:902
def : Pat<(fma (fneg v4f64:$A), v4f64:$C, v4f64:$B),
- (QVFNMSUB $A, $B, $C)>;
+ (QVFNMSUB $A, $C, $B)>;
def : Pat<(fma v4f64:$A, (fneg v4f64:$C), v4f64:$B),
----------------
This is not right as we didn't respect the nsz flag here. But it is not relative with this patch. D76585 miss to handle the QPX. @qiucf
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D78986/new/
https://reviews.llvm.org/D78986
More information about the llvm-commits
mailing list