compute hexorder/atom question

Dear Lammps community,

I want to calculate the orientational order of a 2d system with compute hexorder/atom command. The number of neighbors to be accounted in the calculations is controled by
the parameter nnn. As it is written in the manual if I set nnn NULL then all neighbors up to the cut-off distance are taken into account. (The optional keyword nnn defines the number of nearest neighbors used to calculate qn. The default value is 6. If the value is NULL, then all neighbors up to the distance cutoff are used.)

As a test case I take a triangular lattice with the lattice spacing about 1.0. If I set the nnn =6 then I obtain correct values of the order parameter:

compute 1 all hexorder/atom cutoff 1.1 nnn 6

ITEM: TIMESTEP
0
ITEM: NUMBER OF ATOMS
20000
ITEM: BOX BOUNDS pp pp pp
0.0000000000000000e+00 1.0745699318235420e+02
0.0000000000000000e+00 1.8612097182041992e+02
-1.0745699318235420e-01 1.0745699318235420e-01
ITEM: ATOMS id c_1[1] c_1[2]
1 1 1.11022e-16
2 1 0
3 1 -1.85037e-17
4 1 0
5 1 0
6 1 0

However, if I set nnn NULL the order parameter is not calculated:

compute 1 all hexorder/atom cutoff 1.1 nnn NULL

ITEM: TIMESTEP
0
ITEM: NUMBER OF ATOMS
20000
ITEM: BOX BOUNDS pp pp pp
0.0000000000000000e+00 1.0745699318235420e+02
0.0000000000000000e+00 1.8612097182041992e+02
-1.0745699318235420e-01 1.0745699318235420e-01
ITEM: ATOMS id c_1[1] c_1[2]
1 inf nan
2 inf nan
3 inf nan
4 inf nan
5 inf nan
6 inf nan

Does anyone know what is the problem?

Any help is appreciated.

Sincerely,
Boris

Perhaps Aidan (cc’d) can provide some insight. he is listed as the author of this compute.

axel.