[Pw_forum] File closing errors when using NEB image parallelization

Lorenzo Paulatto paulatto at sissa.it
Tue Jun 16 17:41:08 CEST 2009


In data 16 giugno 2009 alle ore 15:10:54, J. J. Ramsey <jjr19 at uakron.edu>  
ha scritto:
> Image              PC                Routine            Line         
> pw.x               0000000000609818  stop_run_          63     
> stop_run.f90
> pw.x               00000000004331A5  MAIN__             85    pwscf.f90

> While PWscf was compiled with debugging info and the "-traceback"  
> option, I was still compiling against my cluster's already compiled  
> OpenMPI, which should explain why "Unknown" is repeated so much above.

Not all of the "unknown", actually but what's deeper I can deduct from  
what is there.

I cannot say it for certain, bu it looks suspiciously like a race  
condition, possibly related to the kind of filesystem on which the  
wavefunctions directory (outdir, and consequently ~/tmp, per default).
If I had to bet I would say you are using ~/tmp as the default directory,  
and it is a normal NFS share.

Anyway, it is indeed annoying but should be harmless, as long as it only  
happens at the end of the run. I could try and "fix" it now, but it will  
take some time to check for side effects.

best regards

-- 
Lorenzo Paulatto
SISSA  &  DEMOCRITOS (Trieste)
phone: +39 040 3787 511
skype: paulatz
www:   http://people.sissa.it/~paulatto/

     *** save italian brains ***

  http://saveitalianbrains.wordpress.com/


More information about the Pw_forum mailing list