Simon Pilgrim de3605f56b
[TargetLowering] expandFP_TO_UINT - improve fp16 support ...
As discussed on D53794, for float types with ranges smaller than the destination integer type, then we should be able to just use a regular FP_TO_SINT opcode.
I thought we'd need to provide MSA test cases for very small integer types as well (fp16 -> i8 etc.), but it turns out that promotion will kick in so they're unnecessary.
Differential Revision: https://reviews.llvm.org/D54703
llvm-svn: 347251
2018-11-19 19:16:13 +00:00
2018-11-09 18:04:34 +00:00
2018-06-20 22:40:12 +00:00
2018-10-29 20:10:42 +00:00
2018-11-07 11:45:43 +00:00
2018-10-10 14:15:52 +00:00
2018-10-30 09:07:22 +00:00
2018-11-05 14:37:41 +00:00
2018-09-19 18:46:29 +00:00
2018-06-20 19:59:58 +00:00
2018-11-19 19:16:13 +00:00
2018-10-27 15:00:38 +00:00
2018-05-16 08:58:26 +00:00
2018-06-20 22:40:12 +00:00
2018-07-04 19:26:31 +00:00
2018-11-07 06:57:03 +00:00
2018-11-07 06:57:03 +00:00
2018-11-07 06:57:03 +00:00
2018-07-13 15:03:24 +00:00
2018-05-01 13:06:49 +00:00
2018-05-22 13:24:38 +00:00
2018-06-29 16:37:16 +00:00
2018-10-03 22:02:23 +00:00
2018-08-29 14:07:14 +00:00
2018-07-28 00:27:25 +00:00
2018-07-26 10:59:35 +00:00
2018-07-23 23:19:53 +00:00
2016-09-01 14:53:53 +00:00
2018-07-18 14:11:22 +00:00
2018-09-11 15:32:47 +00:00
2018-11-08 19:14:20 +00:00
2018-04-27 13:30:27 +00:00
2018-06-01 10:07:10 +00:00
2011-12-19 20:24:28 +00:00
2018-05-16 08:58:26 +00:00
2018-08-29 14:53:55 +00:00
2018-10-03 22:02:23 +00:00
2018-06-19 16:05:44 +00:00
2018-07-05 09:27:05 +00:00
2018-08-29 14:54:01 +00:00
2018-10-16 08:27:28 +00:00
2018-11-01 10:10:42 +00:00
2018-11-01 13:57:54 +00:00
2018-06-11 16:50:28 +00:00
2018-10-03 22:02:23 +00:00
2018-10-10 14:15:52 +00:00
2018-10-10 14:15:52 +00:00
2018-05-09 02:40:45 +00:00
2018-05-09 02:40:45 +00:00
2018-08-29 14:07:14 +00:00
2018-07-26 10:59:35 +00:00
2018-07-21 16:16:03 +00:00
2018-09-11 09:57:25 +00:00
de3605f56b