It works fine for a simple Lennard Jones test, see attached files.
Energy difference is 0.002468 for a displacement of -0.0001, average
force is about -2.46. Would be even more accurate if the displacement
was smaller. Can you be more specific on the system you are trying to
simulate? Compute group/group doesn't work with all potentials...
Stan
in.lj (259 Bytes)
data.gg_test (301 Bytes)
log.lammps_0.0 (1.78 KB)
log.lammps_0.0001 (1.78 KB)