Rigid dimer SRD example loses atoms

Hi all,

I am interested in SRD solvents and have been experimenting with the examples provided by LAMMPS. The “SRD diffusion demo - dimer particles” (ASPHERE/dimer) comes close to my use case. However, when I increase the number of steps (originally 1.0e5), the simulation loses atoms after 2.05e5 steps (ERROR: Lost atoms: original 17353 current 17349 (src/thermo.cpp:482)).

Is this expected behavior due to the simplicity of the example? If that is the case, I would be very grateful for any advice on how to avoid these errors in simulations involving rigid particles and an SRD solvent.

(patch_28Mar2023_update1 on Debian 11 bullseye)

The LAMMPS examples are set up to showcase how the syntax of commands work, not necessarily to showcase best practices or conservative settings. The runs are usually too short for meaningful results and some aggressive settings may be used.

It is not unexpected. Lost atoms errors can have many reasons and when you are setting up a new system, you will have to validate all settings and make physically sound choices. In this particular case I would try to reduce the timestep to half and double the number of timesteps in the run command and see if you get a stable trajectory.

1 Like