Long range interaction in LAMMPS

Dear Lammps users and developers,
Is it possible to have long-range interaction in LAMMPS (1/r type interaction) for all the pairs (no pairwise cutoff) and Repulsive LJ interaction cut of at (2^(1/6)\sigma).?
How one go about it when you have a periodic system and a system which is confined by walls.
Best Regards
Viswas

Dear Lammps users and developers,
Is it possible to have long-range interaction in LAMMPS (1/r type interaction) for all the pairs (no pairwise cutoff) and Repulsive LJ interaction cut of at (2^(1/6)\sigma).?

the second one is easy to answer. yes. you can define the LJ cutoff for each pair of atom types, and thus make it repulsive-only for all pairs.

LAMMPS uses ewald summation or similar methods (PPPM, MSM) to do computation of full coulomb interactions by computing them in part in real space and the rest in reciprocal space. if you don’t want to benefit from the improved performance of this approach, and do explicit real space calculations only, you can benefit from the fact, that LAMMPS is not subject to minimum image conventions, and thus can use a cutoff larger than half of the extend of the simulation cell, however this is limited by the communication cutoff requirements and the fact, that LAMMPS uses domain decomposition parallelization.

How one go about it when you have a periodic system and a system which is confined by walls.

for confined non-periodic systems i.e., using a cutoff-only approach with a sufficiently long cutoff will give the exact result.

axel.