This very much depends on the application, of course. As my earlier posts discuss, in Python multi-processing is the only way you can really parallelize cpu-bound computations. Threads are not an option.

I am impressed by the brevity-and-clarity of your demo. As always!

Currently prudently starting the rewrite in python+wxpython of an application that may (or may not) be cpu-intensive but certainly cpu-bound, I’d like to try to use an effective way to harness the power of the current cpu(s) without diving into to highly complex synchronisation code.

I definitely hate the idea of sharing code on a multi-thread basis since the focus of this app is definitely not performance.

However since 3.x should be available soon within wxpython, concurrent.futures will be as well.

This application does not requires massively parallel code but could possibly take advantage of a clear cut between a UI process and a calculation one!

I would definitely like the idea of using your code, at least for a start, as a an entry point. Would that make sense in your opinion?

Great post!

I don’t really have an opinion about actors since I haven’t used them. I tried to stick to the standard lib, and ventured outside only to Twisted.

Thank you very much,

Matěj