This chapter gives you an overview of parallel programming architectures and programming models. These concepts are useful for inexperienced programmers who have approached parallel programming techniques for the first time. This chapter can be a basic reference for the experienced programmers. The dual characterization of parallel systems is also presented in this chapter. The first characterization is based on the architecture of the system and the second characterization is based on parallel programming paradigms. Parallel programming will always be a challenge for programmers. This programming-based approach is further described in this chapter, when we present the design procedure of a parallel program. The chapter ends with a brief introduction of the Python programming language. The characteristics of the language, ease of use and learning, and extensibility and richness of software libraries and applications make Python a valuable tool for any application and also, of course, for parallel computing. In the final part of the chapter, the concepts of threads and processes are introduced in relation to their use in the language. A typical way to solve a problem of a large-size is to divide it into smaller and independent parts in order to solve all the pieces simultaneously. A parallel program is intended for a program that uses this approach, that is, the use of multiple processors working together on a common task. Each processor works on its section (the independent part) of the problem. Furthermore, a data information exchange between processors could take place during the computation. Nowadays, many software applications require more computing power. One way to achieve this is to increase the clock speed of the processor or to increase the number of processing cores on the chip. Improving the clock speed increases the heat dissipation, thereby decreasing the performance per watt and moreover, this requires special equipment for cooling. Increasing the number of cores seems to be a feasible solution, as power consumption and dissipation are way under the limit and there is no significant gain in performance.

To address this problem, computer hardware vendors decided to adopt multi-core architectures, which are single chips that contain two or more processors (cores). On the other hand, the GPU manufactures also introduced hardware architectures based on multiple computing cores. In fact, today's computers are almost always present in multiple and heterogeneous computing units, each formed by a variable number of cores, for example, the most common multi-core architectures.

Therefore, it became essential for us to take advantage of the computational resources available, to adopt programming paradigms, techniques, and instruments of parallel computing.