[gmx-users] pbc atom

[chris.neale at utoronto.ca]

velocities are not taken from the .gro as far as I know. You can get  
them from the .trr if you also use a .edr.

If you want to check, then do a run for only 1 ps and save your .edr  
every timestep. Then run g_energy and look at the temperature.

I am not sure that a constant energy simulation in vacuum is a good  
idea. When I say the initial forces become the initial velocities, I  
mean this: you start with no thermal noise, your system is momentarily  
at zero kinetic energy. After the first time step, you get some  
velocities from the initial forces. It is quite likely that these will  
be concerted motions more than random oscillations, especially if you  
are using the pull code. But the kinetic energies are still quite  
small, so your thermostat scales them up massively and you get a  
"flying ice cube" problem (again, read about this please).

I think that you will get farther if you try to test some of these  
yourself: e.g. if you think that you get initial velocities, then  
think of a way to test if you really do (here you have the method I  
suggested above, but it's just an analogy).

Chris.

-- original message --

Thanks for all your help. I am actually considering removing the
thermostat all together seeing as my system is quite small.
I have also started using gen_vel = yes (no matter how much I think my
system is in equilibrium). One quick question (perhaps its silly). When
you say the initial forces become the initial velocities; I thought the
initial velocities were taken from the velocities in the initial
configuration (taking into account that I do have velocities in the
initial configurations from previous simulations).

Cheers

Gavin