[GRASS-dev] Re: multithreading v.surf.bspline and others with OpenMP

[Sören Gebbert]

Hi,

2011/12/18 Hamish <hamish_b at yahoo.com>:
> Sören wrote:
>>> The GRASS ATLAS wrapper is and example for such
>>> an approach. ATLAS can be used, but in case it
>>> is not installed, the default GRASS implementation
>>> is used.
> Hamish:
>> > Oh, I did not know that was there. We can work on
>> > adding it to trunk's ./configure next.
> Sören:
>> We can do, but the IMHO the ATLAS wrapper is not in
>> use by any module, except the library test module.
>
> the question is: if we had support for it would it
> be used? since we already have BLAS for low-level
> stuff and LAPACK for mid-level stuff, and the gmath
> and gpde libraries, and the ccmath library built in,
> it starts to feel a little crowded.  But if it is
> the right tool for the right job I would not want
> to deny someone to easily use it.
>
> what advantage does ATLAS bring to the table?

ATLAS is the C-Implementation of BLAS and parts of LAPACK.[1]

The advantages of ATLAS over the BLAS/LAPACK Fortran wrapper in GRASS
is that its supports
the C-style matrix layout used by the gmath library and that no
Fortran to C conversion is needed.
As far as i know the BLAS/LAPACK wrapper is not used in GRASS and IMHO
with the presents of ATLAS
and ccmath obsolete.

Best regards
Soeren

>
>
> Hamish
>
> ps- does any one know if the equivalent of the
> "jobs" and "wait" shell commands exist for
> launching modules from python?

[1] http://math-atlas.sourceforge.net/