Introduction

In the previous chapter, we saw how to use threads to implement concurrent applications. This section will examine the process-based approach. In particular, the focus is on two libraries: the Python multiprocessing module and the Python mpi4py module.

The Python multiprocessing library, which is part of the standard library of the language, implements the shared memory programming paradigm, that is, the programming of a system that consists of one or more processors that have access to a common memory.

The Python library mpi4py implements the programming paradigm called message passing. It is expected that there are no shared resources (and this is also called shared nothing) and that all communications take place through the messages that are exchanged between the processes.

For these features, it is in contrast with the techniques of communication that provide memory sharing and the use of lock or similar mechanisms to achieve mutual exclusion. In a message passing code, the...