[gmx-users] how to remove an energy contribution in gromacs?

[Berk Hess]

>From: "Jones de Andrade" <johannesrs at gmail.com>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: "Discussion list for GROMACS users" <gmx-users at gromacs.org>
>Subject: Re: [gmx-users] how to remove an energy contribution in gromacs?
>Date: Mon, 6 Aug 2007 11:13:44 -0300
>
>Hi both.
>
>@ Mark: thanks mark. Have to admit that I haven't thought about the
>exclusions possibility. Specially because I was looking for a simple way of
>doing it without removing the dynamics related to these forces from the
>simulation itself. So, changing the nexcl looked a bit easier to do.
>
>@ Berk:
>
> > This seems the right approach to me.
> >
> > I don't know what you are doing exactly.
> > But are you aware that the averages in g_energy and in the log file
> > are using every integration step, whereas with rerun you only get values
> > for every frame in your trajectory?
> > I think the averages of mdrun -rerun are still incorrect, you need to 
>use
> > the average of the values in the energy.xvg file.
>
>
>  Ok, let me see if I understood: This means that, if I simulated 
>originally
>with the .mdp saying for the system to save the trajectory every 10 time
>steps, and even with this fact still mentioned in the .mdp file for the
>"mdrun -rerun" job, it would barelly mean that the energies would be 
>divided
>by a factor of 10, and the "oscilations" are due to the reduction in the
>statistics? Is it correct?

Yes.
The factor will not be exactly 10 though, since there are some end effects.
With may frames it will be very close to 10 though.

>
>So, the proper way of doing if is not just keep on saying on the .mdp file
>that it's saved every 10 time steps, but rather more or I just get the
>results and multiply them by 10, or make the rerun informing a 10 times
>larger time-step?

You can not fix the problems caused by this "bug".
For the average values of the rerun you should just run g_energy
and extract the average from the energy.xvg file using for instance 
g_analyze.

Berk.

_________________________________________________________________
Hotmail en Messenger on the move http://www.msn.nl/services