Commit Graph

4 Commits

Author SHA1 Message Date
Craig Topper a74e3088df [AVX-512] Remove patterns from the other VBLENDM instructions. They are all redundant with masked move instructions.
We should probably teach the two address instruction pass to turn masked moves into BLENDM when its beneficial to the register allocator.

llvm-svn: 291371
2017-01-07 22:20:34 +00:00
Igor Breger 64cfd3a442 [AVX512] Fix BLENDM lowering patterns. Operands should be swapped to match SELECT behavior.
Use BLENDM instead of masked move instruction.

Differential Revision: http://reviews.llvm.org/D21001

llvm-svn: 272763
2016-06-15 07:30:38 +00:00
Igor Breger 484bace21b re-generate the tests using the update_llc_test_checks.py script
llvm-svn: 272643
2016-06-14 07:05:10 +00:00
Elena Demikhovsky b2b901c607 AVX-512: fixed a bug in lowering VSELECT for 512-bit vector
https://llvm.org/bugs/show_bug.cgi?id=23634

llvm-svn: 238195
2015-05-26 11:32:39 +00:00