[gmx-users] Re: temperature

[Mark Abraham]

On 17/04/2012 7:32 AM, Roy wrote:
> Mark,
>
> According to the manual, the temperature of a group should always be
> calculated from the total KE of the group (see eq 3.13 in the manual), which
> presumably includes  contributions from the translational motion of the
> center of mass of the group, plus the rotational motion around the center of
> mass of the group, plus any random non-rigid motion within the group.
> However, according to appendix D.72 for g_traj, the option –ot plots the
> temperature of each group calculated only from the rigid body translational
> velocity of the center of mass of the group, since the command –ot implies
> –com. Aren’t these two definitions of temperature incompatible, since the
> –ot definition ignores two terms (rotational and random)?

While g_traj -h says that g_traj -ot implies g_traj -com, I don't think 
this means that the temperature is necessarily calculated from the 
velocity of only the COM. Rather, I would expect that -ot calculates the 
temperature from the full KE and simply reports the velocity for the COM.

You can test this with g_traj on a .gro file for a small group of atoms 
for which you can construct a zero COM motion and non-zero full KE.

Mark