Bug in calculation of Distance Restraints energy by gmx energy
I have noticed different values of DistRest energy given by gmx energy (on multiple machines). So I tested it on notebook and found that it depends on numbers of MPI, openMP (even with single MPI). More testing with only 2 dist. rest. showed me that some of dist rest were counted for multiple times. Similar problem has a D.R. viol. calculation.
I am using Gromacs 5.1.1., however I tested it with versions 5.0.5 (GPU, MPI) and 4.6.1 and observed different dist. rest. energies. OS Fedora, Debian.
Hope, it will help!