In some cases when synaptic weights are re-allocated after a load_network_state() operation, the pointers to synaptic weights stored in the list of synapses to monitor in WeightMonitor might be out of date. The network simulation will most likely segfault right after the onset of the simulation before completing the first time-step.
A simple fix would be to store the relative offset of weights inside the data array instead of a pointer.
In some cases when synaptic weights are re-allocated after a load_network_state() operation, the pointers to synaptic weights stored in the list of synapses to monitor in WeightMonitor might be out of date. The network simulation will most likely segfault right after the onset of the simulation before completing the first time-step. A simple fix would be to store the relative offset of weights inside the data array instead of a pointer.