[gmx-users] beware gen-vel=no

[chris.neale at utoronto.ca]

Regarding gen-vel=no as discussed here:
http://lists.gromacs.org/pipermail/gmx-users/2011-May/061151.html

I would caution against the general use of gen-vel=no and then  
coupling to a temperature of 300 K with the Berendsen thermostat. One  
problem that can arise is concerted unfolding. Temperature is random  
velocity, but with gen-vel=no and a structure that is just slightly  
too compact, you will have a slight net force that gets massively  
scaled up by the coupling algorithm and results, not in true  
temperature, but in large scale coordinated motion.

It's just something to look out for. gen-vel=no should not be a  
problem with the sd algorithm or with some type of position restrained  
equilibration procedure or, of course, if loading in velocities from  
some cpt or trr file.

Chris.