projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 659e1b7) | patch
[X86] Prefer fpext(splat(X)) to splat(fpext(x)).
authorFreddy Ye <freddy.ye@intel.com>
Mon, 16 Jan 2023 14:16:02 +0000 (22:16 +0800)
committerFreddy Ye <freddy.ye@intel.com>
Mon, 16 Jan 2023 14:48:30 +0000 (22:48 +0800)
commitdb711f79ef57c96782f3e5093ee9b79509a317a4
tree757b8495f30096d025516c1c48bcdc0a749846fetree | snapshot
parent659e1b7bb09eba04ce90cc072279263b61c63156commit | diff
[X86] Prefer fpext(splat(X)) to splat(fpext(x)).

This patch is to fix regression of D122875. X86 has fpext instructions
supporting rmb form, which takes advantage of fpext(fplat(X)) than
splat(fpext(X)).

Reviewed By: RKSimon, skan

Differential Revision: https://reviews.llvm.org/D141657
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/lib/Target/X86/X86ISelLowering.h diff | blob | history
llvm/test/CodeGen/X86/prefer-fpext-splat.ll diff | blob | history
Domain: System / Toolchain;