[Pw_forum] GIAPW LSF problem?
Axel Kohlmeyer
akohlmey at cmm.chem.upenn.edu
Fri Apr 25 18:07:24 CEST 2008
On Fri, 25 Apr 2008, Charles wrote:
charles,
the even better solution would be to get rid
of LSF. it is creating a lot of problems
with typical "cluster usage" _and_ costs a lot
of money. PBS/Torque+Maui don't cost you zip
and work much better. unfortunately, there are
too many sysadmins around that don't understand
that something that costs a lot can be worse
than something that costs nothing... :-(
cheers,
axel.
CC> Thanks Lorenzo,
CC>
CC> I found a easy solution.
CC>
CC> Basically it is to use -i tag of LSF rather than use -in tag for pw,
CC> bands, pp, etc. This also works for gipaw.
CC>
CC> Maybe this can be added as another note in the user guide for people
CC> have to struggle with the input redirection.
CC>
CC> This works for me with mpi+lsf system
CC>
CC> Charles
CC>
CC> Lorenzo Paulatto wrote:
CC> > On Ven, Aprile 25, 2008 05:41, Charles Chen wrote:
CC> >> So, the gipaw.x does not read the input file.
CC> >
CC> > There is no -in nor -input option in gipaw.x, at the moment the only way
CC> > to feed it the input is via standard input.
CC> >
CC> >> Is there something I need to specify when I make gipaw.x, so that it can
CC> >> read the input file with -in tag?
CC> >
CC> > Yes, you have to program it ;-).
CC> >
CC> >> Or this is a problem purely related to LSF?
CC> >
CC> > Maybe you can still use mpi + lsf whith input redirection. It usually
CC> > works, but you need the input file on all nodes. (You can create it in the
CC> > input script using the "cat << eof ... eof" trick as in the examples)
CC> >
CC> > regards
CC> >
CC>
CC> _______________________________________________
CC> Pw_forum mailing list
CC> Pw_forum at pwscf.org
CC> http://www.democritos.it/mailman/listinfo/pw_forum
CC>
--
=======================================================================
Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu http://www.cmm.upenn.edu
Center for Molecular Modeling -- University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.
More information about the Pw_forum
mailing list