[gmx-users] comm-grps

[Erik Lindahl]

Bert de Groot wrote:

>Hi all,
>
>I've got a question concerning com-grps. There are two groups
>in my system of which I would like to keep the center of mass
>translation to zero. So I've defined those groups (plus a rest
>group) in make_ndx, and hoped that the following .mdp setup
>should do the job:
>
>
>; mode for center of mass motion removal = 
>comm-mode                = Linear
>; number of steps for center of mass motion removal = 
>nstcomm                  = 1
>; group(s) for center of mass motion removal = 
>comm-grps                = around_1695 around_907 others
>
>
>In the simulation, however, I see a significant c.o.m. motion
>of all of the groups, that seems to be exactly coupled
>with the deformation of the box size (which is quite large, 
>on purpose). Is there a way to freeze the c.o.m. motion in an 
>absolute sense (so irrespective of any changes in the box dimensions)?
>
>cheers,
>
>Bert
>
Hi Bert,

This might have to do with the way we scale the coordinates when the box 
is changed
in Berendsen pressure coupling. Do you see it with Parrinello-Rahman 
coupling too?

Cheers,

Erik