[CP2K:3240] status of hybrid OpenMP+MPI version

[Axel]

hi urban,

> Actually, I also get crashes with OpenMPI--but not with MPICH2.  I
> recompiled OpenMPI 1.4.3 with MPI user thread support, and I recompiled
> BLACS and ScaLAPACK with it.  As soon as I use more than 1 thread (only
>
i think that one went away after defining   -D__HAS_NO_OMP_3

> tried popt) it crashes in a BLACS call.
>
> I'll also give the OpenMPI 1.5 development branch a try.
>

hmm... i have this nagging feeling that there may be a race condition 
somewhere
and that it just doesn't show with MPICH.

with experimenting some more, i see that the "degree of badness" in the 
total
energies, for example, is rather low with using only 2 or 3 threads but gets 
large
with 6 or more.

on the other hand, i am _very_ impressed how well the threading 
parallelization
scales (with the .ssmp binary), even to a very large thread count (i have a 
few
48 core maschines to test on).

i hope there is an answer somewhere as to why hybrid parallelization does 
crash
with OpenMPI...

cheers,
    axel.

Regards,
> Urban
>
> > 
> > thanks for your help,
> >      axel.
> > 
> > 
> >         just make sure we're on the same page. the input(s) that
> >         are giving me the problems are the water benchmark inputs.
> >         in tests/QS/benchmark/
> >         
> >         
> >         the cp2k.?smp binaries work fine for OMP_NUM_THREADS=1,
> >         but as soon as i enable more than one thread with cp2k.psmp
> >         and multiple MPI tasks, i get NaNs or random energies.
> >         cp2k.ssmp appears to be working for both intel 11.1 and
> >         gfortran 4.4.4
> >         
> >         
> >         thanks,
> >             axel.
> >         
> >         
> >                 Cheers,
> >                 Urban.
> >                 
> >                 > 
> >                 > in any case, i'll give it a try.
> >                 > 
> >                 > 
> >                 > thanks,
> >                 >      axel.
> >                 > 
> >                 > -- 
> >                 > You received this message because you are subscribed
> >                 to the Google
> >                 > Groups "cp2k" group.
> >                 > To post to this group, send email to
> >                 cp... at googlegroups.com.
> >                 > To unsubscribe from this group, send email to cp2k
> >                 > +uns... at googlegroups.com.
> >                 > For more options, visit this group at
> >                 > http://groups.google.com/group/cp2k?hl=en.
> >                 
> >                 
> > 
> > -- 
> > You received this message because you are subscribed to the Google
> > Groups "cp2k" group.
> > To post to this group, send email to cp... at googlegroups.com.
> > To unsubscribe from this group, send email to cp2k
> > +unsu... at googlegroups.com.
> > For more options, visit this group at
> > http://groups.google.com/group/cp2k?hl=en.
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20110512/8ddce781/attachment.htm>