Thanks Peter! I imagine the fault lines should be strongly dependent on the region format (for now only squares, easily extendable to rectangles). The code should be generalizable to 3D without difficulty, I should try that too.

Hey Szabolcs!

I hadn't obtained that error personally, so I played around until I was able to repeat it. It seems createDisks wasn't evaluated in your nb, it should just be a matter of evaluating the initialization cells.

The intention was part scientific experiment and part learning experience, to try to keep it 100% in WL (otherwise I would indeed do it in C++). Thanks for the package suggestion though, I'll definitely use it.

As for the performance, the average simulation takes a couple minutes, but is highly sensitive on initialization parameters (play around, I want to know how to make it better).