[gmx-users] two questions on multiple simulations and stochastic dynamics

[Nguyen Hoang Phuong]

> Hi, all:
>     I have two questions.
> 1) I want to perform multiple simulation on one system and run them in
> parallel with gmx3.2.1. What I did is:
> grompp -v -f md0.mdp -c pr.gro -p x.top -o topol0.tpr
> grompp -v -f md1.mdp -c pr.gro -p x.top -o topol1.tpr
> grompp -v -f md2.mdp -c pr.gro -p x.top -o topol2.tpr
> grompp -v -f md3.mdp -c pr.gro -p x.top -o topol3.tpr
>
> Then I sunmit this job to a cluster which had one processor on each node
> by using "qsub myjob". myjob is a script:
> #PBS -l nodes=4,walltime=240:00:00
> cd working_directory/
> lamboot
> mpirun -np 4 mdrun -multi -np 4 -s topol.tpr
>
> It did produce four trajectories. But all four simulations were running
> on the first processor and leave the rest three nodes idle. Is it possible
> to let one subjob run on one of the four processors or I should do
> grompp -np 4 -v -f md0.mdp -c pr.gro -p x.top -o topol0.tpr ...?
> Any hints will be sincerely appreciated!

you can use the script file:

j=-1

while [ $j -lt 3 ] ; do

let j=j+1

grompp_d -v -f md${j}.mdp -c pr.gro -p x.top -t ttt${j}.trr -o topol.tpr -np 1

done

mpirun -np ${ncpu}  mdrun_d -np ${ncpu} -multi -s topol.tpr  -e md.edr -o md.trr -c md.gro

you will get 4 files md0.trr,...md3.trr

Hope that help.



>
> 2) I am using the stochastic integrator in gmx3.0.5. I notice that there
> is only the inverse friction constant used in mdp, does this mean that
> the friction force here is only related to the friction coefficient of
> solvent molecules without considering the fraction of solute surface
> exposed to the solvent? In GROMOS96 package, the friction force is related
> to both. Since the friction coefficient of water is 91 (ps-1), what
> parameter should be set for tau_t as inverse friction constant in
> stochastic dynamics of biomolecules? Any help will be sincerely
> appreciated!
>
>
>
> --
> Best regards
>
> Yours sincerely
>
> Hao Fan :-)
>
> -------------------------------------------------------------------------
> Drs. Hao Fan                            email F.Hao at chem.rug.nl
> Lab. of Biophysical Chemistry
> University of Groningen
> Nijenborgh 4                            9747 AG GRONINGEN
> The Netherlands
>
>
> Please don't call me on Friday due to the group meeting.
> -------------------------------------------------------------------------
>
> _______________________________________________
> 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.
>