Dear lammps users
I have 6164 atoms in my system. I used the L-J and coulombic potential for the system. The lammps version is lammps-30Jul16 compiled with the mpi makefile on the cluster. I used 12 cores for my simulation and the time step is 1fs. 41 hours have passed, but only 170 ps simulation is finished. Is this a right efficiency for lammps? Is there other ways to improve efficiency. Part of my input is included.
units real
atom_style full
dimension 3
boundary p p s
neighbor 2.0 bin
neigh_modify delay 5
read_data water.data
#L-J potentials
pair_style hybrid lj/cut/coul/cut 15 lj/cut 15
pair_coeff 1 1 lj/cut 0.05570 3.4
pair_coeff 2 2 lj/cut/coul/cut 0.00000 0.000
pair_coeff 3 3 lj/cut/coul/cut 0.15535 3.166
pair_coeff 1 3 lj/cut 0.09302 3.283
pair_coeff 2 3 lj/cut/coul/cut 0.00000 0.000
pair_coeff 1 2 lj/cut 0.00000 0.000
bond_style harmonic
bond_coeff 1 554.1349 1.0
angle_style harmonic
angle_coeff 1 45.7696 109.47
timestep 1
Someone suggested the hybrid style is not necessary here(I will correct it for next simulation.). Is this related to the efficiency? If this is not the right efficiency? Where I did wrong?
Fan Li