Atom count is inconsistent


I am running a simulation on a system in varying salt concentrations. I have already successfully ran the system in water and 3M NaCl. My higher concentration system reaches its final values for minimization within a few steps, then remains there for the rest of minimization. If I try to run a defrost I get the error “Atom count is inconsistent, cannot write restart file” and several values have very high exponents (energy, pressure, enthalpy). I am re-neighboring every 2 steps and have already tried increasing changing the box bounds and rebuilding the system from scratch. I have visualized my input and no atoms seem to be overlapping. I have attached my data file, minimization, and defrost files. Any advice would be greatly appreciated!

Thank you,
Olivia Cracchiolo

minimization_output.txt (13.7 KB)

minimization.txt (1.07 KB)

defrost_output.txt (3.14 KB)

DefNPT.txt (1.98 KB)

6na_nma_data.txt (329 KB)


things go bad before that. the indication is the excessively low pair energy. that is and indication that two atoms overlap. possibly one of the water hydrogen atoms with a Cl- anion.
you can confirm this, but using write_data or write_dump after minimization and then do a ‘delete_atoms overlap 0.5 all all’ followed by another write_data/write_dump. comparing the two files should indicate the atom that is overlapping and then finding what is close to it should be straightforward. you may need to increase the force constants for water bond and angle interaction to avoid having the water too much deformed by the ions during minimization (since you cannot use fix shake, as needed for this kind of water potential).

the restarting is doomed, since the close contact will just “explode” those atoms out of the system and thus cause the error.