Реалізація паралельного алгоритму бульбашкового сортування

Abstract

Розглянуто дослідження алгоритму бульбашкового сортування та можливості його паралельної реалізації з використанням бібліотеки OpenMP і стандартних засобів багатопотоковості мови C++ (std::thread).

The study of the bubble sort algorithm and the possibilities of its parallel implementation using the OpenMP library and standard C++ multithreading support (std::thread) is considered. The basic concepts of sorting, the principles of the bubble sort algorithm, and the features of parallel computing are analyzed. The choice of development tools and approaches for implementing the parallel algorithm is justified. A sequential and parallel implementations of the bubble sort algorithm are developed, and their performance is experimentally evaluated on data sets of various sizes. The experimental results show that the use of multithreading significantly improves performance when sorting large data arrays with an optimal number of threads, while further increasing the number of threads does not always lead to performance gains due to synchronization overhead and memory bandwidth limitations. The obtained results confirm the effectiveness of parallel computing for accelerating sorting algorithms.