Технічні науки English Відкритий доступ
Comparison of computational complexity of programs using parallel programming in Python and C++
Флягін Владислав Костянтинович, Олійник Олена Володимирівна
Міжнародний науковий журнал «Інтернаука», № 18, грудень 2021
Анотація
Nowadays, computational tasks are everywhere and amount of data rises every year, so using parallel programming become more and more important. In cases like neural networks computation, which can be easily divided into separate processes, due to all operations are matrixes and we can perform functions simultaneously, regardless of the order. Particularly at the moment, when GPUs have become much more accessible than a number of years ago, using GPUs significantly ameliorate computational time. In this article we are considering C++ OpenMP [1] library, which supports multi-platform shared-memory parallel programming in C/C++ and Fortran, also defines a portable, scalable model with a simple and flexible interface for developing parallel applications on platforms from the desktop to the supercomputer, in compare to Python multiprocessing module. We picked a Pi evaluating problem as a computational task, because Pi is one of the most important world constants, which evaluation does matter. As a result of the article, we stated that using C++ and “pragma omp” directive is better than Python and it’s multiprocessing module due to a Python internal processes. We propose you to follow our path and try this code on your own. Each line of the following code is written on our own and is accessible on GitHub [2] to fully restore our results and dive a bit dipper into this topic on your own.
Ключові слова: parallel programming; OpenMP; multiprocessing
Як цитувати
Fliahin V., Oliynyk O. Comparison of computational complexity of programs using parallel programming in Python and C++ // International scientific journal "Internauka". — 2021. — №18.
@article{internauka7781,
author = {Флягін, Владислав Костянтинович and Олійник, Олена Володимирівна},
title = {Comparison of computational complexity of programs using parallel programming in Python and C++},
journal = {Міжнародний науковий журнал «Інтернаука»},
year = {2021},
number = {18},
url = {https://www.inter-nauka.com/issues/2021/18/7781}
}
TY - JOUR AU - Флягін, Владислав Костянтинович AU - Олійник, Олена Володимирівна TI - Comparison of computational complexity of programs using parallel programming in Python and C++ JO - Міжнародний науковий журнал «Інтернаука» PY - 2021 IS - 18 SN - 2520-2057 UR - https://www.inter-nauka.com/issues/2021/18/7781 LA - en ER -