[gmx-users] mdrun -multi

[David]

On Mon, 2003-12-01 at 17:07, Nguyen Hoang Phuong wrote:
> > >
> > > I would like to run mdrun with the option -multi (multiple simulations in
> > > paralell). Suppose that I have 4 nodes (8cpus), is it possible to run 8
> > > simulations in parallel on these 8cpus?
> >
> > AFAIK there is one, and only one!, sensible reason for using mdrun -multi
> > and that is when you have access to a supercomputer with rather strict
> > policies on the queues. That is, many supercomputer administrators require
> > that only parallel (and thus multi-processor) jobs can be run. If, in stead,
> > you have a (large) number of individual simulations that you want to run
> > you can use mdrun -multi to bypass this queuing policy by making the system
> > think you are running a paralell job, while you actually run multiple
> > independent simulations.
> >
That is not the only reason, and certainly not a reason for implementing
it! It was intended for weird parameter optimizations.

> > In all other cases, IMHO, you should simply start multiple separate instances
> > of mdrun!
> 
> thanks Anton. Here we have a linux cluster with many queues but each
> queues has only 2 cpus. If I use mdrun -np 4 -multi then four jobs will be
> submitted to one queue and  running on 2cpus. Are there possibilities
> (some trick or modify the code) to distribute these 4 jobs on 4
> cpus (2 queues)? Of course without starting multiple separate
> instances of mdrun.
In principle it should start as many jobs as you specify on the number of CPUs that 
your queuing system gives you. If you want to perform four jobs, you
would just submit two jobs of two jobs to the queue.


> Phuong
> 
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.
-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assist. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://xray.bmc.uu.se/~spoel
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++