compute orientorder/atom

Dear all,

I would like to look at ice crystallisation and I would like to use the compute orientorder to get information on the ordering of the molecules. My simulation runs fine without dumping this compute. But as soon as this compute is performed I get an error on the neighbor list;

Neighbor list overflow, boost neigh_modify one (../npair_full_bin.cpp:119)

I tried different settings for the compute (adding "cutoff 3.0").

My input file is:

units metal
atom_style full

read_data melt.data

velocity all create 150.0 4928459

the error message is telling you what to do.
the orientorder compute needs a “full” neighbor list instead of the “half” neighbor list of the pair style.
at typical water densities and a (large) cutoff of 15.4 angstrom, you are running over the default limit for neighbor list entries per atom (2000).
for a system at density 1.0 g/cm^3 you would have around 2500 neighbors in a full neighbor lists. that overflows the 2000 of the default setting.
please see the discussion in the documentation for the neigh_modify command.

axel.

Thank you, Axel.
I did not realize the “full”/“half” difference.

Herma