To be an interpreted language, Python is fast, and if speed is critical, it easily interfaces with extensions written in faster languages, such as C or C++. A common way of using Python is to use it for the high-level logic of a program; the Python interpreter is written in C and is known as CPython. The interpreter translates the Python code in an intermediate language called Python bytecode, which is analogous to an assembly language, but contains a high level of instruction. While a Python program runs, the so-called evaluation loop translates Python bytecode into machine-specific operations. The use of interpreter has advantages in code programming and debugging, but the speed of a program could be a problem. A first solution is provided by third-party packages, where a programmer writes a C module and then imports it from Python. Another solution is the use of a Just-in-Time Python compiler, which is an alternative to CPython, for example, the PyPy implementation optimizes code generation and the speed of a Python program. In this book, we will examine a third approach to the problem; in fact, Python provides ad hoc modules that could benefit from parallelism. The description of many of these modules, in which the parallel programming paradigm falls, will be discussed in subsequent chapters.

However, in this chapter, we will introduce the two fundamental concepts of threads and processes and how they are addressed in the Python programming language.