Problem with Point Process Neuron Network Simulation

[mszuromi]

Describe the bug
I have a bug when constructing a network of point process neurons. In jupyter notebook, I cannot run simulations longer than 117.1 ms (exactly) within a single cell. I can run longer simulations, but I have to break it up into multiple cells of 100 ms runs.

To Reproduce
This occurs for a wide variety networks of point process neurons. I have used both the pp_psc_delta neuron model, and I have built my own simplified neuron model through NESTML. The problem is identical regardless of network size, connection rules, synapse model, whether the NESTML module is created with an accompanying synapse, etc. The problem is immediately alleviated if I substitute the point process for an iaf neuron model with the same network structure and rules.

Expected behavior
I expect the simulation to run in a comparable amount of time within a single notebook cell as when split between several cells.

Screenshots
have attached a pdf of my script that works (with simulation time = 100 ms per cell). If the second "nest.Simulate(100)" is uncommented, the simulation will never finish (I have let it run for 24+hrs).

Desktop/Environment (please complete the following information):

• Computer: M2 Macbook Pro
• OS: iOS
• Browser: Chrome & Safari
• Shell: zsh
• Python-Version: 3.7, 3.10, & 3.11
• Jupyter Notebook-Version: 6.4 & 7.2
• NEST-Version: 3.7
• Installation: Docker & Anaconda

Additional context
This problem occurs for multiple versions of python, jupyter notebook, NEST installation method, and browser. I have also assigned multiple local threads to the run. Memory and CPU usage looks normal while running the script.

Network w NESTML.pdf