x86 f16: implement rounding support#1380
Conversation
|
Some errors from Microsoft Visual Studio 2015 version 19.00.24241.7 and Visual Studio 2017 version 19.16.27035: https://ci.appveyor.com/project/nemequ/simde/builds/53462546/job/ya2shs5yojpisa31#L4688 |
9e8035c to
234fb23
Compare
|
Hmm, this is weird. It's rounding |
|
The only difference from 19.14 to 19.20 (which gives the expected answer) seems to be some spills: https://godbolt.org/z/Y59G5vK65 |
|
Actually, it appears they changed |
|
I just removed the test since signed nan handling already differs between targets |
8d1733d to
5d6d7fd
Compare
Head branch was pushed to by a user without write access
|
Looks like the Azure CI pipeline started timing out? |
|
@TheNumbat I will rebase this PR once #1384 is merged |
Assumes `_MM_FROUND_CUR_DIRECTION` is always half-to-even, which isn't strictly correct, but is already assumed by various other functions.
|
Thank you @TheNumbat ! |
|
np! also, for context, I'm now using simde to run https://github.com/janestreet/ocaml_simd on arm 👍 |
Very cool, thank you for letting us know! Do you want to send them a PR to incorporate SIMDe directly? |
|
It's already incorporated internally (I work at Jane Street) - should appear in the repo soon-ish |
Huzzah! Can you add an entry at https://github.com/simd-everywhere/simde/wiki/Projects-Using-SIMDe ? 💞 |
Makes
_mm_cvtps_ph/_mm256_cvtps_phrespect the rounding mode argument. Assumes_MM_FROUND_CUR_DIRECTIONis always half-to-even, which isn't strictly correct, but is already assumed by various other functions.Enables the rounding mode tests and adds a few more. There should be no behavior changes for the round-nearest and round-current cases.