[gmx-developers] dispersion correction / exploding box

[Berk Hess]

Rainer Böckmann wrote:

>Dear David,
>
>the dispersion energy is quite high, starting at 5*10**4 and decreasing with
>increasing box volume to ~3*10**4 within 70ps (d=1.5nm between protein and box).
>A similar simulation with d=1.2nm with stable volume shows a dispersion energy
>of -1*10**4.
>
>  
>
Do you really mean that the dispersion correction energy is positive??

I have recently changed some stuff in the dispersion correction and
am currently working on it again.

But all my runs use some CVS version with dispersion correction and
give the correct energies and densities.

Berk.