Summary
We looked at a couple of technologies that we can exploit to make our Python code run faster and, in some cases, use multiple CPUs in our computers. One of these is the use of multiple threads, and the other is the use of multiple processes. Both are supported natively by the Python standard library.
We looked at three modules: threading, for developing multithreaded applications, multiprocessing, for developing process-based parallelism, and concurrent.futures, which provides a high-level asynchronous interface to both.
As far as parallelism goes, these three modules are not the only ones that exist in Python land. Other packages implement their own parallel strategies internally, freeing programmers from doing so themselves. Probably, the best known of these is NumPy, the de-facto standard Python package for array and matrix manipulations. Depending on the BLAS library that it is compiled against, NumPy is able to use multiple threads to speed up complex operations (for example,...
