The ABINIT Users Mailing List ( CLOSED )

[Mauro Sgroi <maurosgroi@yahoo.it>]

Dear Yann,
thanks a lot for the advices.
I will optimize the code for the different cpus.
I've only one further question: is it correct that
excecuting the parallel calculation I have to put the
input and pseudo files locally on each PC?
If I have no local copies of that files, the code
exits complaining about missing input file.
Thanks a lot and best wishes,
Mauro.

--- Yann Pouillon <yann.pouillon@ehu.es> ha scritto:

> Hi,
> 
> Le Fri, 6 Jul 2007 09:17:16 +0200,
> maurosgroi@yahoo.it a écrit :
> 
> > I've a cluster composed by 2 pentium IV and 1 AMD
> Athlon64.
> > Should I optimize the abinit code for a generic
> architecture and use
> > exactly the same excutable on all CPus (for
> example supplying
> > -march=generic to gfortran) or I can optimise the
> code for each
> > architecture and use excutables optimised on each
> PC?
> 
> Here we have a very similar cluster (well ... with
> many more processors,
> but still the same kind), and we use two optimised
> versions of the code, each
> of them located in a directory which can be seen
> only from one type of
> processor. The directories are mounted so that they
> appear with the same path
> to the running jobs, then the users do not have to
> care about where they are
> performing their calculations.
> 
> I don't know if this is the best solution, but it is
> working pretty well.
> 
> Best regards,
> 
> Yann.
> 
> -- 
> Yann Pouillon               European Theoretical
> Spectroscopy Facility (ETSF)
> Tel: (+34) 943 01 83 94     Centro Joxe Mari Korta,
> Avenida de Tolosa, 72
> Fax: (+34) 943 01 83 90     20018 Donostia-San
> Sebastián (Gipuzkoa)
> Web: http://www.etsf.es/    España
> 
> 



      ___________________________________ 
L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail: 
http://it.docs.yahoo.com/nowyoucan.html