This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: math tests: reduced precision on Opteron 6272
- From: Aurelien Jarno <aurelien at aurel32 dot net>
- To: Joseph Myers <joseph at codesourcery dot com>
- Cc: libc-alpha at sourceware dot org
- Date: Thu, 30 Jun 2016 00:31:59 +0200
- Subject: Re: math tests: reduced precision on Opteron 6272
- Authentication-results: sourceware.org; auth=none
- References: <20160629211245 dot GA3688 at aurel32 dot net> <alpine dot DEB dot 2 dot 20 dot 1606292126490 dot 18822 at digraph dot polyomino dot org dot uk>
On 2016-06-29 21:28, Joseph Myers wrote:
> On Wed, 29 Jun 2016, Aurelien Jarno wrote:
>
> > Would it make sense to regenerate the libm-test-ulps on this CPU?
>
> Yes. (Long-term if the MPC performance issues for inverse trig and
> hyperbolic functions get fixed it will be possible to move these tests to
> auto-libm-test-in and probably generate more tests to make the ulps more
> accurately reflect the maximum errors of the functions, in which case
> there may be less dependence on the processor.)
Thanks for the details, I have just done that.
Aurelien
--
Aurelien Jarno GPG: 4096R/1DDD8C9B
aurelien@aurel32.net http://www.aurel32.net