[gmx-users] Vacuum simulation freezes the system

[Martin Ulmschneider]

Hi,

I have found that during a vacuum simulation my peptide (gly-gly-gly) 
slowly freezes when i use the Berendsen thermostat. I have checked that 
this applies to all proteins in vacuum, for example with the nmr1 
tutorial supplied with the gromacs distribution (just change the number 
of steps to 100000 or so).

My question is this: How can i perform a realistic simulation in vaccuum 
that does not freeze? I.e. what do i need to change in the mdp-file?

Best wishes,
Martin