make check fails

[Geert Jan BEX]

Dear all,

Thanks for your suggestions, I'll try tuning the -fp-model option.

Best regards, -gjb-

On Mon, Jan 19, 2015 at 6:43 PM, Torbjörn Granlund <tg at gmplib.org> wrote:

> Marc Glisse <marc.glisse at inria.fr> writes:
>
>   Yes, Intel disables denormals by default, without changing the value
>   of denorm_min to reflect this fact. Their bug. We should probably add
>   something to configure to use a suitable -fp-model option for Intel's
>   compiler, just in case...
>
> I suppose we could try adding this option to configure.ac, but we cannot
> regression test it, since icc if non-free (as in both beer and freedom).
>
> --
> Torbjörn
> Please encrypt, key id 0xC8601622
>



-- 
- dr. Geert Jan BEX
- HPC analyst/consultant Hasselt University & University of Leuven
- email:                geertjan.bex at uhasselt.be
- homepage:         http://alpha.uhasselt.be/~gjb/
- PGP public key: http://alpha.uhasselt.be/~gjb/pgp.txt
---------------------------------------------------------