projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: eee554d) | patch
[x86, AVX] optimize masked loads with constant masks
authorSanjay Patel <spatel@rotateright.com>
Wed, 9 Mar 2016 22:12:08 +0000 (22:12 +0000)
committerSanjay Patel <spatel@rotateright.com>
Wed, 9 Mar 2016 22:12:08 +0000 (22:12 +0000)
commit4a8dd89128eb563d50ac9dd2c9f8017fc0cfc495
tree8115ff1605c1384b63da181b674072e71b7e05cftree | snapshot
parenteee554d86ed2407623421d3aa3bac1f8e0d9e2aecommit | diff
[x86, AVX] optimize masked loads with constant masks

Instead of a variable-blend instruction, form a blend with immediate because those are always cheaper.

Differential Revision: http://reviews.llvm.org/D17899

llvm-svn: 263067
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/test/CodeGen/X86/masked_memop.ll diff | blob | history
Domain: System / Toolchain;