[gmx-users] Extending simulation not work: always re-start from the beginning

[Justin Lemkul]

On 6/8/15 9:19 AM, Zhang, Cheng wrote:
> Dear Gromacs expert,
>
> I was trying to extend my simulations. Relevant files can be found at
>
> https://copy.com/c8dDMnC0df2gbig9
>
>
> To prepare the "md_0_1.tpr" file, I set the following in the "md_50ns.mdp" so that it can run 50 ns in total.
>
>
> nsteps = 25000000
>
> dt    = 0.002
>
>
> Then I submit the same job for 5 times with the following command:
>
>
> gerun mdrun_mpi -deffnm md_0_1.tpr -cpi md_0_1.cpt -maxh 30 -append
>
> (Each of the 5 jobs would be hold until the previous one had finished.)
>
>
> However, as I check the 5 xtc files (e.g. RMSD):
>
>
> #md_0_1.tpr.xtc.1#
>
> #md_0_1.tpr.xtc.2#
>
> #md_0_1.tpr.xtc.3#
>
> #md_0_1.tpr.xtc.4#
>
> md_0_1.tpr.xtc
>
>
> all of them start the simulation from 0 to 10000 ps with different fluctuations. As I understand, they all start from the beginning instead of extending the simulation based on its previous one.
>
>
> So can I ask if I have done something incorrect?
>

If you're not modifying the contents of the .tpr file to indicate that you want 
to extend the simulation, mdrun is just going to run the same simulation.  It's 
doing what you have told it to do.  You need to invoke gmx convert-tpr 
(previously called tpbconv).  See the wiki:

http://www.gromacs.org/Documentation/How-tos/Extending_Simulations#Version_4_and_Newer

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================