[gmx-users] Question about center-of-mass removal

[Berk Hess]

>From: Jae Hyun Park <jaepark at uiuc.edu>
>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] Question about center-of-mass removal
>Date: Tue, 11 Jul 2006 00:27:55 -0500
>
>Thank you, Mark.
>
>I have one more question for the confirmation. So, if there
>is other asymmetries (e.g. initially the fluid is at one
>side and the other side is empty, and as time goes on, it
>spreads to the empty side - non-equilibrium process. In flow
>direction there is no periodicty while periodic in other two
>directions), do I have to turn off the center-of-mass
>removal (e.g. nstcomm = 0) ?

Everything depends on if your system has an absolute reference
in space, in Gromacs this can only be done with postion restraints
or freeze groups.
You have not written if this is the case.

If there is no absolute reference, your com is uncoupled and you
should remove the com motion.
If you have an absolute reference you should not remove com motion
as this will influence the sampling in the system.

Fluid flow itself has nothing to do with this. If the fluid flows in one
direction this is because of a force working on it and there should
be an equally large reaction force working on something else
in the opposite direction.

Berk.