projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 48615c1) | patch
[X86] Don't mark SEXTLOAD v4i8->v4i64 and v8i8->v8i64 as custom under vector widening...
authorCraig Topper <craig.topper@intel.com>
Sat, 29 Dec 2018 01:17:11 +0000 (01:17 +0000)
committerCraig Topper <craig.topper@intel.com>
Sat, 29 Dec 2018 01:17:11 +0000 (01:17 +0000)
commit0a6cec6f9f1bd11147eb76d8d0b4df9d40bb873c
treecec9b3a0c24ac760d704834ac51ad603fb3a6ba2tree | snapshot
parent48615c17866a50a36b7ea6f0e435c7e9b77df702commit | diff
[X86] Don't mark SEXTLOAD v4i8->v4i64 and v8i8->v8i64 as custom under vector widening legalization.

This was tricking us into making these operations and then letting them get scalarized later. But I can't prove that the scalarized version is actually better.

llvm-svn: 350141
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/test/CodeGen/X86/vector-sext-widen.ll diff | blob | history
Domain: System / Toolchain;