Compute cluster/atom with rigid body


I want to analyze the clusters during the simulation and tried compute cluster/atom. It turns out atoms in a rigid body would be marked as different clusters even though they are within the cutoff. I read through the manual and it seems a rerun is required to do the analysis. However I did not find an example on combining compute cluster/atom with rerun.
Should I do something like

compute cluster/atoms 1.5
dump clusters all custom 100000 cluster.dat c_clusters
rerun trj.lammpstrj dump id mol type x y z

Any suggestion would be appreciated!


Compute cluster/atom does not know anything about rigid bodies. But you also must not use neigh_modify exclude to exclude intra-cluster interactions, since that will mean those pairs of atoms are removed from the neighbor list and that means that compute cluster/atom doesn’t see them either.

If all clusters have a different molecule ID, you may want to check out compute fragment/atom instead.

Thank you very much I think I got it!