You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@arrow.apache.org by "Antoine Pitrou (Jira)" <ji...@apache.org> on 2019/09/18 11:22:00 UTC

[jira] [Commented] (ARROW-6436) [C++] vendor a half precision floating point library

    [ https://issues.apache.org/jira/browse/ARROW-6436?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16932318#comment-16932318 ] 

Antoine Pitrou commented on ARROW-6436:
---------------------------------------

See also ARROW-3802. Numpy has dedicated float16 routines that we could reuse. Given that it's Numpy it's probably well-maintained.

There may be other projects around.

> [C++] vendor a half precision floating point library
> ----------------------------------------------------
>
>                 Key: ARROW-6436
>                 URL: https://issues.apache.org/jira/browse/ARROW-6436
>             Project: Apache Arrow
>          Issue Type: Improvement
>          Components: C++
>            Reporter: Benjamin Kietzman
>            Priority: Major
>
> Clang and GCC provide _Float16 and there are numerous polyfills which can emulate a 16 bit float for other platforms. This would fill a hole in the kernels and other code which don't currently support HALF_FLOAT.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)