[gmx-users] (no subject)

[Mark Abraham]

----- Original Message -----
From: Nilesh Dhumal <ndhumal at andrew.cmu.edu>
Date: Tuesday, October 26, 2010 10:50
Subject: Re: [gmx-users] (no subject)
To: Discussion list for GROMACS users <gmx-users at gromacs.org>

> I run a test simulation for -rerun. I didn't change the topology file.
> 
> grompp -f md.mdp  -c  solvent-bmi-pf6-128.pdb  -
> p  solvent-bmi-pf6-128.top
>  -o 3.tpr
> mpirun -machinefile cp -np 8 mdrun -s 3.tpr -o 3.trr   
> -c
> solvent-bmi-pf6-128.pdb -e 3.edr -g 3.log
> 
> with -rerun
> grompp -f md.mdp  -c  solvent-bmi-pf6-128.pdb  -
> p  solvent-bmi-pf6-128.top
>  -o 6.tpr
> mpirun -machinefile cp -np 8 mdrun -s 6.tpr -o 6.trr -rerun 
> 3.trr  -c
> solvent-bmi-pf6-128.pdb -e 6.edr -g 6.log
> 
> I calculate the total energy by
> g_energy -f 3.edr -o 3.xvg
> g_energy -f 6.edr -o 6.xvg
> 
> The total energy varies between +- 30.00 KJ/mol.
> It should be constant since I using same topology file and trajectory.
> Why the total energy is not constant.

Those .tpr should be identical - but you can check that with gmxcheck. Reruns do neighbour-searching every step, whereas your normal simulation followed the nstlist setting. That's part of why my earlier advice suggested doing reruns for each simulation you wish to compare. You should be able to get good/better agreement for steps where nstlist directed neighbour-searching in the original run. Also, whether or not constraints have been applied (and when!) could influence the energies to about this degree. I don't recall the details here.

Even once you've removed all algorithm-specific sources of difference, there are other sources of non-reproducibility, such as the assignment of particles to DD cells. Your original mdrun probably used dynamic load-balancing, and that cannot be reproduced in the DD used by the rerun. (Or indeed by a repeat of your original mdrun!) Setting -dlb no in the original simulation might be enough to get agreement here, or maybe mdrun -reprod will be required.

Mark

> NIlesh
> 
> 
> 
> 
> On Thu, October 21, 2010 10:24 am, Mark Abraham wrote:
> > On 22/10/2010 1:17 AM, Nilesh Dhumal wrote:
> >
> >> I am doing solvation dynamics for my system.
> >> I have system with diatomic (PA---NE)solute surrounded by water
> >> molecules.
> >>
> >> I want to run simulation with two differcent cases.
> >> 1. PA charge=0 and NE charge=0 : No charge on solute
> >> 2. PA charge=+1 and NE charge=-1 : Charge on solute
> >>
> >>
> >> I want to calculate the energy at each step keeping the solvent
> >> configration same.
> >>
> >> IF I start a simulation with no charge on solute (case1), I 
> have the
> >> energy for 1 step. I want to calculate the energy with charge 
> on solute
> >> (case 2) with same configration water molecules.
> >> Each step  I want to calculate the energy with and 
> without charge on
> >> solute since the configration of solvent will be same for 
> that step.
> >>
> >> I was thinking two make two topologies file with charge and 
> with out
> >> charge on solute. I don't know how to use them simultaneously 
> during the
> >>  simulation.
> >
> > Well, you don't use them simultaneously. You run a simulation on
> > whatever you think will generate a relevant conformational 
> ensemble. Then
> > you want to use mdrun -rerun twice on the resulting 
> trajectory, using .tpr
> > files based on .top files corresponding to the two cases in 
> order to
> > create your comparison.
> >
> > Mark
> > --
> > gmx-users mailing list    gmx-users at gromacs.org
> > http://lists.gromacs.org/mailman/listinfo/gmx-users
> > Please search the archive at
> > http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> > Please don't post (un)subscribe requests to the list. Use the
> > www interface or send it to gmx-users-request at gromacs.org. 
> Can't post? Read
> > http://www.gromacs.org/Support/Mailing_Lists
> >
> >
> >
> 
> 
> 
> -- 
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at 
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20101026/5a92fcb6/attachment.html>