[gmx-users] NVE simulations

[Mehmet Suezen]

"David L. Bostick" wrote:
> It all depends on that initial configuration and velocities for an NVE
> simulation. 

Hi David,

Isn't true that, "in principle" equilibrium configuration should not
depend
on initial conditions?  If not than I'd say the system is not in
equilibrium state
and instead of changing ensemble will extent NVT  simulation or decrease
the integration time step before NVE run. One measure for an equlibrium
state  would be checking Maxwell-Boltmann velocity distribution or
Boltmann H-function where it decays to a constant value if system
approach to equilibrium. 

Cheers,

Mehmet