This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: automatically updating ulps -- any review needed ?
- From: Andreas Jaeger <aj at suse dot com>
- To: libc-alpha at sourceware dot org
- Date: Mon, 06 Jan 2014 07:33:05 +0100
- Subject: Re: automatically updating ulps -- any review needed ?
- Authentication-results: sourceware.org; auth=none
- References: <201401051711 dot 28152 dot vapier at gentoo dot org>
On 01/05/2014 11:11 PM, Mike Frysinger wrote:
> the ia64 code hasn't had its ulps file regenerated in a long time. i used the
> helper target for regenerating things, and now i'm looking at the diff. but i
> have no idea how to evaluate what these differences mean :). most are easy --
> new baselines for tests that didn't have any info recorded previously.
>
> i tried reading math/README.libm-test, but it seems like some base knowledge
> is assumed.
>
> for example, i don't know what to make of things like:
> Function: "yn":
> double: 3
> -float: 2
> +float: 3
> idouble: 3
> -ifloat: 2
> +ifloat: 3
> ildouble: 2
> ldouble: 2
We have ULPs generated per test case - and then one maximum per
function, the above is the one per function. The per function one is
shown in the manual.
> or am i thinking too hard about this and i should just commit it and forget ?
for cases like yn, just ignore it, an ULP of 2 or 3 is fine. you should
check that no functions that have to be exact like sqrt are in the list.
If in doubt, just send it here and have Joseph or myself look over it,
Andreas
--
Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg)
GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126