Shed Skin restricted-Python-to-C++ compiler, feedback?

I like how it’s the Othello example that’s winning the most here :wink:

I’m kinda curious about how it compares against Cython and numba now.

What is the y axis on this graph measuring?

Mojo picked Mandelbrot for their 35x --zeros-- speed up claim too. It’s embarrassingly parallel and vectorisable.

Is Shedskin still generating single threaded code in this example? It’s not also effectively throwing more cores at it, and comparing the speed up to a naive base Python case which only used one of the hundreds of CPUs assumed to be available?