[gmx-users] Potential Energy problem

[Mark Abraham]

On 11/05/2010 10:45 PM, Carla Jamous wrote:
> Thank you Mark,
>
> I'm running gromacs 4.0.3., so I don't know if I should submit a bugzilla.

There's nothing in the release notes that I could see that looks like 
it's been fixed. 
http://www.gromacs.org/About_Gromacs/Release_Notes/Revisions_in_4.0

Since there's been so many bugfixes, you should probably upgrade to 
4.0.7 anyway. If so, and you can replicate this issue, please submit a 
bugzilla.

> Please Do you know another way to restart my MD run?

The problem is that your checkpoint time step (54ps) is greater than the 
range in your .tpr (0-12ps). If you -extend your .tpr far enough, or 
that can't arise. A better solution is probably to use init_step and 
tinit in the .mdp file (see manual 7.3) so that eventually you can make 
simulation that looks contiguous.

I'd be concerned whether the original bizarre heating was indicative of 
a serious problem with the simulation or code, however.

Mark

> Thanks,
> Carla
>
> On Tue, May 11, 2010 at 1:00 PM, Mark Abraham <Mark.Abraham at anu.edu.au
> <mailto:Mark.Abraham at anu.edu.au>> wrote:
>
>     On 11/05/2010 7:55 PM, Carla Jamous wrote:
>
>         Hi everyone,
>
>         please I encountered a problem of potential energy rising suddenly
>         during my simulation.
>         Someone advised me to take the previous conformation of my system,
>         center it in the box, minimize it & then restart my MD
>         simulation, by
>         taking the previous velocities of this conformation.
>         So I did the centering and the minimization.
>
>         Usually, I used to do the following for my MD run.
>         tpbconv -s .tpr -extend 6000 -o .tpr
>         mdrun -v -s .tpr -cpi state_.cpt -cpo state_.cpt -o .trr -c .gro
>         -x .xtc
>         -e .edr -g .log
>
>         However, this time I need to go from the minimized structure.
>         So in order to get a .tpr file of my minimized structure (.gro), I
>         created a .mdp file with gen_vel=no and did:
>
>         grompp -f .mdp -c minimized.gro -p .top -o minimized.tpr
>
>         After that, I tried the following:
>
>         tpbconv -s minimized.tpr -extend 6000 -o .tpr
>         mdrun -v -s .tpr -cpi state_.cpt -cpo state_.cpt -o .trr -c .gro
>         -x .xtc
>         -e .edr -g .log
>
>         and got this error message:
>
>         *WARNING: This run will generate roughly 3410368410012306432 Mb
>         of data
>
>         starting mdrun 'Protein in water'
>         6 steps,  12.0 ps (continuing from step 27,  54.0 ps).
>
>
>     That looks like a bug. The run time from your .mdp+tpbconv is 12ps,
>     but the .cpt is already past that and GROMACS is using integer
>     arithmetic inappropriately to get a huge number of steps (thus large
>     output volume estimate) but no actual steps occur.
>
>     If you're running GROMACS 4.0.7, please submit a bugzilla.
>
>     Mark
>     --
>     gmx-users mailing list gmx-users at gromacs.org
>     <mailto:gmx-users at gromacs.org>
>     http://lists.gromacs.org/mailman/listinfo/gmx-users
>     Please search the archive at http://www.gromacs.org/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
>     <mailto:gmx-users-request at gromacs.org>.
>     Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>
>