Re: [ESPResSo] Visualisation of MD trajectories - assigning of angle and dihedral parameters - rdf over more configs ?

[Axel Arnold]

Hi!

On Tuesday 06 April 2010 00:48:36 Marek Maly wrote:
> #2
> Is in actual Espresso possible to set angle and dihedral parameters just
> using
> particle types of bonded particles similarly like in atomistic forcefields
> are predefined these parameters
> according to the atom types of the bonded triplet or quaternion of atoms ?
>
> Let say that I have particles of type 0 and 1 in my system. It is possible
> to define "globally" angle parameters
> for all the relevant  combinations of bonded particle types:
>
> 0-0-0, 0-0-1, 0-1-0 , 1-0-1, 1-1-1 (for angle parameter setting)
>
> or
>
> 0-0-0-0, 0-0-0-1 ... (for dihedral parameters)

No. Angular and dihedral potentials are only available as bonded potentials, 
which are specified on a per-particle basis. Of course, you can write a Tcl 
procedure which checks the particle types and adds the necessary bonds.

> #3
> Is there available exmple of Espresso calculation where the
>
> analyze <rdf>
>
> command is used to calculate radial distribution function averaged over
> previously stored configurations ?

Look into the "First steps" section of the Users' Guide. There is a small 
tutorial, in which among other things the rdf is calculated. Whether 
configurations were stored before or are generated on the fly doesn't matter, 
if you read them back.

> Is it really necessary to store configurations in files to calculate <rdf>.

It is not necessary to store the configurations, as I said above - whether 
configurations are first stored and then read back, or are analyzed at 
production, doesn't matter. In fact, the analzye rdf command works always on 
the internal particle coordinates of the integrator. Reading back for analysis 
just means setting these internal coordinates.

> I mean is it possible to calculate <rdf> during simulation from
> let say last N loops of the INTEGRATE "for" cycle (of course without my
> own tcl coding).

Without Tcl coding, Espresso won't do anything, so the answer is no. You will 
need to write your own averaging, just as in the Users' guide.

Regards,
Axel

-- 
Dr. Axel Arnold Tel: +49 711 685 63593
ICP, Universität Stuttgart      Email: address@hidden
Pfaffenwaldring 27
70569 Stuttgart, Germany