This is the mail archive of the
ecos-discuss@sourceware.org
mailing list for the eCos project.
Re: gcc 4.1.2, some testcases fail when built with optimizations
- From: Gary Thomas <gary at mlbassoc dot com>
- To: Hans Rosenfeld <rosenfeld at grumpf dot hope-2000 dot org>
- Cc: ecos-discuss at ecos dot sourceware dot org
- Date: Sun, 17 Jun 2007 13:58:06 -0600
- Subject: Re: [ECOS] gcc 4.1.2, some testcases fail when built with optimizations
- References: <20070617195610.GA29929@grumpf.hope-2000.org>
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hans Rosenfeld wrote:
> I noticed that some testcases fail when built with -O2 optimizations.
> The failing tests are cos, ldexp, log, log10, pow, sin and tan from libm
> and the cpuload test.
>
> I think the cpuload test fails because the compiler is removing some
> code that has no result and no side effects (like busy delay loops), so I
> don't care about it that much.
>
> The libm tests fail because the tested functions really produce wrong
> results, like cos() returning a value > 1. For the cos() case I traced
> this to __ieee754_rem_pio2(), which seems to fail completely for some
> values. I have no idea what exactly is going on here.
>
> Does anyone else experience such problems?
What's the architecture (ppc, mips, etc)? I run 4.1.2 here all
the time with no troubles.
- --
- ------------------------------------------------------------
Gary Thomas | Consulting for the
MLB Associates | Embedded world
- ------------------------------------------------------------
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org
iD8DBQFGdZJOmaKbSsQGV8ARAtknAJ93VQL6G4fEQ2LmGajdXx8du4z5AACeN86d
4+jEEvf1canuLpIH7ncTxxE=
=DtUW
-----END PGP SIGNATURE-----
--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss