Re: [AMBER] single longer simulation Vs multiple shorter simulation

From: Jason Swails <>
Date: Fri, 2 Dec 2011 09:42:57 -0500

On Fri, Dec 2, 2011 at 2:18 AM, Bala subramanian

> Amber users,
> I would like to know the merit and demerit of doing a multiple short
> simulations instead of a single longer simulation.
> I am working on protein complex which is known to undergo significant
> conformational changes in complex form. So we expect the changes to happen
> in longer time scale and decided to do longer simulations (NPT ensemble
> with langevin coupling). However i would like to know if there is any
> advantage of doing multiple shorter simulations in such cases.

It depends on what you mean by multiple shorter simulations. If they're
independent, then you won't see anything that occurs on a timescale longer
than your simulation time. If each one is a restart from the previous
simulation, then it shouldn't make much of a difference.

One advantage to running short simulations (and just restarting them) is
shorter wall-clock times, which often enough bumps your priority up in a
scheduler and allows you to run faster. Also, schedulers (such as PBS)
allow you to submit multiple jobs at the same time and have them released
sequentially after the one before it finishes (so that it doesn't run
before), which moves the held jobs up in the queue while the first is


> Thanks,
> Bala
> _______________________________________________
> AMBER mailing list

Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
AMBER mailing list
Received on Fri Dec 02 2011 - 07:00:04 PST
Custom Search