Concurrency with Python, Twisted, and Flex

Article by Bruce Eckel.

"(...) to make use of more than one CPU we need true parallelism, where multiple CPUs are working on the problem at the same time. To accomplish this in Python in this article, we create multiple operating system processes, one for each solver (worker) (The master only requires occasional CPU cycles and it shares the CPUs to do this). A separate Python interpreter is started in each process, and each process has its own memory so there is no possibility of collision as there is with threads, which makes multi-process parallelism much easier and safer than thread programming."