energy_data.nlogn_use -- select non-bond list generation algorithm

Before calculating dynamic non-bonded restraints, MODELLER determines which of the several routines is most appropriate and efficient for calculating the non-bonded atom pairs list. The user can use this variable to influence the selection, of either a straightforward ${\cal O}(n^2)$ search or a cell-based algorithm which has $n \log n$ dependency of CPU time versus size $n$ . The latter algorithm is used when the maximal difference in residue indices of the atoms in the current dynamic restraints is larger than energy_data.nlogn_use or when the box size for this algorithm would have to be larger than 8.