This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug math/20033] [x86_64] vectorized math function don't call the __finite versions
- From: "andrew.n.senkevich at gmail dot com" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sourceware dot org
- Date: Tue, 26 Jul 2016 16:20:23 +0000
- Subject: [Bug math/20033] [x86_64] vectorized math function don't call the __finite versions
- Auto-submitted: auto-generated
- References: <bug-20033-131@http.sourceware.org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=20033
--- Comment #2 from Andrew Senkevich <andrew.n.senkevich at gmail dot com> ---
(In reply to Julian Taylor from comment #1)
> I disagree with this. The scalar fallback allows me to have equal results to
> scalar math for non finite inputs which makes them much more useful for
> general purpose libraries.
>
> Currently also with vector math pow(nan, 0) will give me 1 as I would with
> scalar math. If you call pow_finite this property is probably not retained.
Yes, finite version of pow(nan, 0) returns qNaN instead of 1.0.
And it was a reason to exclude that testcases in the patch -
https://sourceware.org/ml/libc-alpha/2016-07/msg00213.html :)
If current way already used somewhere and it is helpful lets stay it as is?
--
You are receiving this mail because:
You are on the CC list for the bug.