Random initialization of molecules

LAMMPS LAMMPS Mailing List Mirror
1

Hello,

I am trying to initialize 2-atom molecules randomly using create_atom.

molecule mols2 mols2.mol
create_atoms 0 random 10 12345 atom_box mol mols2 383

My mols2 file looks like this:
2-atom molecule

2 atoms
1 bonds
0 angles
0 dihedrals
0 impropers

Types

1 1
2 2

Bonds

1 1 1 2

Coords

1 0 0 0
2 0 0 0.97

However, I am getting the error “Bond atoms # # missing at proc # at step #” error. My suspicion is that when I initialize randomly, the second atom is being initialized too far away from the first atom. Is this what is causing this error? If this is indeed the error, how can I prevent this?

Thank you!

Best,
H. Yao

2

No, I don’t think so, because then you would get an error on timestep 0. More likely is that the atoms are moving too fast for LAMMPS, either because initial velocity is too large or forces are too large. LAMMPS spatial decomposition (even on 1 process) can not keep track of atoms that move more than 1 processor domain between reneighborings. This is not normally an issue, but sometimes very simple or very defective simulations can break it. Monitor velocities every timestep and you will understand the problem better.