[gmx-developers] short range energies

[Mehmet Suezen]

David van der Spoel wrote:
> >as far as I've seen in ns.c in function search_neighbours(), how it puts
> >particles in to list if one choose bham for example? Or, how could one
> >be sure
> >that neighborlist constructed properly?
> I should have thought of this before: if you set an environment variable
> setenv DUMP_NL 1
> it will dump the entire neighbourlist in your md.log file, so you can
> check whether all interactions are there. The format is a bit terse,
> please check the source in wnblist.c
> 

tests reports has show that neighbourlists constructed but somehow
energy is still printed as zero (bham and the new one), it is quite
annoying that lj type is functioning properly but not the others,
somewhat
contradictory because selection path is almost identical for all
short-range potentials in the code, isn't it? 

I've already cheked that correct loops are calling, parameters are in
force.c/ns.c are 
identical. what else could be the problem or any suggestion would be
valuable indeed.
cheerz,

Mehmet