[lammps-users] About fix langevin

Dear all,

I have here a short question regarding the fix langevin command in lammps:

as we know, in fix langevin since "each processor uses the input seed to generate its own unique seed and its own stream of random numbers", the dynamics of the system will not be identical on two runs on different numbers of processors. Then if the two runs are peformed with same numbers of same type processors, will thee dynamics be definitely identical?

And a question from me which could be primitive for scientists in Sandia: Is it possible to have some options for fix langevin like "loop geom" for the velocity command, so that the random number works globally independent of the the # of processors?


1st Q: if you repeat the same run on the same # of procs, it should
be identical.

2nd Q: It could be done in principle, but I don't think it's a good
idea to base the random # on the atom's position. Why do you care,
so long as it's random?