Dear LAMMPS developers and users,
I am modifying LAMMPS codes for my own purpose (I am adding a volume swap capability into the temper command).
It works fine initially, but when the program restarts and run a couple of steps, atom velocities of some replicas suddenly turn to be -nan.
To fix this problem, I need to trace back to see what happened inside the time integration (I did not touch “update->integrate->run(nevery)”).
It would be of great help if you can direct me to the variables/subroutines that store the atom force/acceleration, so that I can proceed with the debugging…
Hopefully I have made myself clear enough. Thank you so much for the help and time.
LC Liu