[NEST Users] Simulation execution time for extended runs

The simulation of C.elegans now is implemented in a multi processing environment. The simulation of each individual object and various observation tools such as visualisation, matplotlib spike data plotting and configuration of the simulation parameters have their own process. An issue has arisen with only one object being simulated in that the execution 'time' for the constant 'simulation step time', increases with subsequent call to pynn.nest.Simulate(self.simulation_step). I presume nest continues the simulation from the previous iteration state. A crude plot suggests that the time increase is linear. Eventually the simulation stalls. I have a recorder connected to each of the 302 nodes and multiple recorders connected to some of the nodes. These use multiprocessor queues and a manged list to transfer data and as far as I can tell they do not overflow. I have disabled the recorders by not connecting them to the neurons and the phenomenon persists. It looks as if the issue is in the nest simulation. I have increased the process priority to -1. It does not help. Also does anyone know of the work of Ramin Hasani at Wien. He seems to have done the same thing and is now using 19 neurons to steer a car. How can I get a copy of the report of work done by him and his students? Peter Mason