| commit | 47ec2c7479de7527c955c93fce1323717c58782c | [log] [tgz] |
|---|---|---|
| author | Ahmed Bougacha <ahmed.bougacha@gmail.com> | Mon Mar 23 17:48:07 2015 +0000 |
| committer | Ahmed Bougacha <ahmed.bougacha@gmail.com> | Mon Mar 23 17:48:07 2015 +0000 |
| tree | 00053954d42ec8a0c2dfb1d58cca8de30f2207ba | |
| parent | affe181b397be2bef404e0158b3a8410795fd716 [diff] |
[CodeGen] Convert double -> __fp16 in one step. Fix the CodeGen so that for types bigger than float, instead of converting to fp16 via the sequence "InTy -> float -> fp16", we perform conversions in just one step. This avoids the double rounding which potentially changes results from a natural IEEE-754 operation. rdar://17594379, rdar://17468714 Differential Revision: http://reviews.llvm.org/D4602 Part of: http://reviews.llvm.org/D8367 llvm-svn: 232968