[Wrf-users] Re: wrf.exe Stops without errors when running wih 2 ...

Maite Merino mmerino at am.ub.es
Wed Feb 4 06:38:34 MST 2009


Hi Hailong,

Thanks for your suggestion, but I'm afraid that's not my case. I 
already work with unlimited stack size. And in my case, I get no error 
at all: wrf with 2 or more procs simply never ends for a rather quick 
simulation that with 1proc finishes ok. Besides, I've run several 
times in the same cluster huge MM5 and CMAQ simulations in paral.lel 
mode with up to 24procs without problems, so it should not be a 
problem regarding the cluster ability to work in paral.lel and/or with 
big simulations.

Best Regards,

Maite Merino


On Tue, 3 Feb 2009, Hailong Wang wrote:

> Hi Maite,
> I had a similar problem when I started running ARW in parallel mode. It runs 
> fine on less than 4 processors but got "segmentation fault" on more than 4 
> ..... My problem was in the "stacksize" setting. We suppose to set "limit 
> stacksize unlimited"
> (I use C shell). In my case, this setting did not work properly in job script 
> file, but has to be in the ".cshrc" file.
> Hope it helps.
>
> Hailong 
> -- 
>
>
>
> Hailong Wang, PhD
>
> CIRES/University of Colorado, and ESRL/NOAA
>
> Mail: 325 Broadway, R/CSD2, Boulder, CO 80305
>
> Tel:  303 497 6615
>
> Fax: 303 497 5822
>
> E-mail: Hailong.Wang at noaa.gov
>
>

________________________________________________________________

    Maria Teresa (Maite) Merino Espasa    Maite.Merino at am.ub.es

    Departament d'Astronomia i Meteorologia
    Facultat de Fisica
    Universitat de Barcelona
    Avinguda Diagonal, 647                Tel: +34 93 403 92 33
    E-08028 Barcelona                     Fax: +34 93 402 11 33
    SPAIN
________________________________________________________________



More information about the Wrf-users mailing list