So after letting my Espresso project lay idle for awhile I've picked up the pieces of a connected Gay-Berne particle interaction simulation. Briefly: I'm in the process of tuning the time stepping to prevent unphysical results. I immediately noticed that after halving the time step the simulation's average kinetic energy per particle actually dropped even though I kept the temperature set point constant. I haven't looked into the source code but this suggests something isn't quite right.
Does anyone have any suggestions as to how I should go about trouble shooting this?