[lammps-users] Evolution of the center of mass of polymers


I have 1000 polymers made up of 200 lennard jones atoms with fene bond
connections. I want to calculate the diffusion coefficient of the
polymers. I can use compute com or compute msd. Compute msd does not
give good statistics as it does not use the values of previous
timesteps so the best way that I see to go about this is to dump the
center of mass coordinates of each polymer and calculate the mean
squared displacement using the com's. However, I am limited to 32
groups which I can use in "compute com". I could dump the coordinates
of each atom and post process to get the polymer center of mass
coordinates but this would mean I will be dumping 200 numbers more in
comparison to just dumping the center of mass of the polymer.

Maybe there are new features in LAMMPS that I don't know to get around
my problem of how to efficiently obtain the center of mass of more
than 32 polymers. Any help or suggestions is very much welcomed.


Check out compute com/molecule and compute msd/molecule.
If you want to do continuous time averaging from a bunch
of t=0 starting points, then I think you do indeed need to post
process that, e.g. from compute com/molecule data.