[gmx-users] DispCorr

[Erik Lindahl]

Hi,

>
>Should the following parameter settings result in equal simulation results if all other parameter are equal:
>
>a) DispCorr           = EnerPres
>   Pcoupl             = Parrinello-Rahman
>   Pcoupltype         = anisotropic
>   tau_p              = 1
>   compressibility    = 6e-5 6e-5 6e-5 6e-5 6e-5 6e-5 
>   ref_ p             = 1 1 1 1 1 1
>

You probably don't want any skewing, so the last three (off-diagonal) 
elements of ref_p should be 0. It won't change
things significantly, though, since we're talking orders of magnitude...

>
>
>b) DispCorr           = Ener
>   Pcoupl             = Parrinello-Rahman
>   Pcoupltype         = anisotropic
>   tau_p              = 1
>   compressibility    = 6e-5 6e-5 6e-5 6e-5 6e-5 6e-5 
>   ref_ p             = 624 624 624 1 1 1
>
>(I calculated the 624 bar acoording to the formular given in the book of Allen & Tildesley.)
>

Yes, but there might be statistical fluctuations due to rounding errors, 
etc., so don't be alarmed if your trajectories aren't identical.
 (The average long-time properties should be , though, provided your 
calculated value is right)

Cheers,

Erik