maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
242,208 Commits 23 Branches 0 Tags 2.2 GiB
Commit Graph

2 Commits

Author SHA1 Message Date
Simon Pilgrim 74b3bfdf71 [InstCombine][X86] Regenerate SSE combine tests as part of setup for D17490 
Regenerated with utils/update_test_checks.py 

llvm-svn: 266731
 2016-04-19 12:56:46 +00:00
Simon Pilgrim 20c607b110 [InstCombine] CVTPH2PS Vector Demanded Elements + Constant Folding 
Improved InstCombine support for CVTPH2PS (F16C half 2 float conversion):

<4 x float> @llvm.x86.vcvtph2ps.128(<8 x i16>) - only uses the bottom 4 i16 elements for the conversion.

Added constant folding support.

Differential Revision: http://reviews.llvm.org/D12731

llvm-svn: 247504
 2015-09-12 13:39:53 +00:00