Re: [ESPResSo] Verlet Lists

[Axel Arnold]

On Monday 19 July 2010 18:20:22 Mikheil Azatov wrote:
> Hi Espresso users,
>
> I wanted to ask question about the structure of Verlet pair lists in
> Espresso.
> According to verlet.h for the non-bonded interactions, the integrator uses
> verlet pair lists which contain all particle pairs with a distance smaller
> than \ref max_range_non_bonded = \ref max_cut_non_bonded +  \ref
> integrate::skin. So does a verlet list contain all the particles or just
> the particles that are involved in the non-bonded interactions?

As you cite it, it contains ALL.

> Also, does Espresso create Verlet list even if there are no non-bonded
> interactions in the simulation?  (i.e just from setting skin parameter). I
> would like to have it because not having non-bonded interactions would save
> some simulation time for me but I still need Espresso to create Verlet list
> to go over checks that In included in forces.c.

No, if there are no non-bonded interactions in the system, the Verlet lists 
are not built. If you have an algorithm that wants to use the verlet lists, it 
should add its cutoff to calc_maximal_cutoff in interaction_data.c, so that 
the Verlet lists are correctly set up. Just playing with the skin would not 
help, since then the code also allows particles to move out of their position 
by skin, until the lists are rebuild, so that you still would not catch all 
pairs.

Many regards,
Axel

-- 
JP Dr. Axel Arnold Tel: +49 711 685 67609
ICP, Universität Stuttgart      Email: address@hidden
Pfaffenwaldring 27
70569 Stuttgart, Germany