I have noticed a very strange behaviour of the current (new) version of EDIP code. The EDIP potential code was changed recently (23 Aug 2012) due to a sign error in one of the force terms. As Daniel Karls noticed “Since, at the point in pair_edip.cpp where the coordination forces are calculated, dr_ij = Ri-Rj, we should still get a minus sign out front, which is missing in the current version of the source.” But at lines 306-308 (of pair_edip.cpp) one can see that dr_ij = Rj-Ri. So, maybe the previous sign convention was correct. Possibly, the problem comes from the fact that at lines 189-191 the relation used by Daniel Karls (dr_ij = Ri-Rj) is used to compute environment coordination.