Dear LAMMPS,
I noticed this in the code for fix rigid: fix_rigid.cpp removes 3N-6 degrees of freedom from each rigid body consisting of two or more atoms. While this is in general true, it is not true for diatomic or otherwise linear molecules. In that case, only 3N-5 dof should be subtracted from each rigid body. This effects the calculation of the temperature and pressure, however temp_modify can be used to account for the discrepancy with dof. Perhaps a line can be added to the fix_rigid documentation?
thanks,
Joyce