GROMACS version: 2019.3
GROMACS modification: No
Here post your question
I sometimes encounter the error that water molecules cannot be settled at a quite late stage (after tens of ns), which usually leads to a fatal error concerning domain decomposition in the end. I think the root cause should not come from the degree of parallelization since the same setting has worked well for similar systems, and if it did, the error should have occurred much earlier from my experience.
As for other common reasons causing simulation to blow up: The energy minimization was done until Fmax <100 kJ/mol with a combination of steep and cg. The NPT equilibration was done in 100 ps with Berendsen barostat and the system density reached plateau after around 30 ps. Then the production run was performed with timestep = 2 fs, V-rescale thermostat tau-t = 0.5 ps, P-R barostat tau-p = 5 ps. The production run had been well for over 40 ns until this error occurred.
In addition, the system is quite large and heterogeneous with a protein of 300 residues surrounded by several polymer chains. I am wondering if this needs special considerations in MD settings.
Appreciate for any suggestion of probable reason or advice of parameter change.
