I was looking at the man page for read_restart and it mentions that fix langevin will not restart and continue exactly. I am assuming that this has to do with an interruption of the random number sequence. Hence the question, if you restart with the same random seed, is there a way to fast forward the random number generator to the same point as when the simulation left off?
Thanks,
Fred Phelan
if you restart with the same random seed, is there a way to fast
forward the random number generator to the same point as when the simulation
left off?
no, not with the RN generator that LAMMPS uses. Few good RN generators
are able to do that, especially if they store significant state.
And knowing how many times it was used
in the preceeding simulation would be a problem as well.
And restarting on different numbers of processors would still be
not be reproducible.
Can you explain why you want to do this?
Steve