Dear all,
I am trying to simulate a simple LJ system. How can I eliminate the
neighbor list and any other tricks used in LAMMPS?
Any help would be appreciated.
Regards,
Ali
Dear all,
I am trying to simulate a simple LJ system. How can I eliminate the
neighbor list and any other tricks used in LAMMPS?
Any help would be appreciated.
Regards,
Ali
Dear all,
I am trying to simulate a simple LJ system. How can I eliminate the
neighbor list and any other tricks used in LAMMPS?
Any help would be appreciated.
ali,
if you don't want the "tricks" in LAMMPS, don't use LAMMPS.
a "simple" LJ code can be written very easily. check out
your favorite text book on MD. basically all of them have
examples on how to do it. mind you the _physics_ will be
the same. the "tricks" in LAMMPS only help to execute it
faster and in parallel.
cheers,
axel.
Axel,
I know the tricks help to execute LAMMPS faster but this is not true all
the time. For example, when the dump command writes the positions in a
file, some atoms go slightly outside the simulation box because the
neighbor lists are rebuilt. It is important to fix this problem for
calculating heat flux by using of Green-Kubo method. I am trying to know
how I can use LAMMPS without neighbor list. Can I set the skin parameters
equal to a large value and set the delay parameters to a large number for
building only a big neighbor list?
Axel,
I know the tricks help to execute LAMMPS faster but this is not true all
the time. For example, when the dump command writes the positions in a
file, some atoms go slightly outside the simulation box because the
neighbor lists are rebuilt. It is important to fix this problem for
calculating heat flux by using of Green-Kubo method. I am trying to know
how I can use LAMMPS without neighbor list. Can I set the skin parameters
equal to a large value and set the delay parameters to a large number for
building only a big neighbor list?
this is a postprocessing problem. just rewrap your coordinates
_after_ the simulation and before your analysis and you are fine.
no need to sacrifice performance.
of course you can also reneighbor in every step. see neigh_modify.
axel.
The dump file has nothing to do with neighbor lists. Neighbor
lists are internal to LAMMPS; you don't see their effect in output.
Steve