Zobrazit v EDS

RESEARCH OF PROGRESSIVE TOOLS OF PARALLEL COMPUTATIONS WITH THE USE OF SIMD ARCHITECTURE.

Uloženo v:
Podrobná bibliografie
Název: RESEARCH OF PROGRESSIVE TOOLS OF PARALLEL COMPUTATIONS WITH THE USE OF SIMD ARCHITECTURE.
Alternate Title: ДОСЛІДЖЕННЯ ПРОГРЕСИВНИХ ЗАСОБІВ ПАРАЛЕЛЬНИХ ОБЧИСЛЕНЬ ІЗ ЗАСТОСУВАННЯМ SIMD АРХІТЕКТУРИ (Ukrainian)
Autoři: Zhulkovskyi, O. O., Zhulkovska, I. I., Vokhmianin, H. Ya., Firsov, O. D., Riabovolenko, V. A.
Zdroj: Informatics & Mathematical Methods in Simulation / Informatika ta Matematičnì Metodi v Modelûvannì; 2023, Vol. 13 Issue 3/4, p228-235, 8p
Témata: COMPUTERS, BIG data, COMPUTER performance, MATRIX multiplications, SOFTWARE development tools, PARALLEL programming, COMPUTER assisted instruction
Abstrakt: The current stage of development of processes and technologies requires continuous improvement of computer hardware performance, efficient use of its resources, processing of large amounts of data and support of the growing requirements of modern information systems. When processing large amounts of data, it is often necessary to use additional effective solutions to speed up information processing in addition to parallel computing. One such approach is to use the SIMD mechanism. The concept of SIMD instructions is a progressive solution for speeding up computations in tasks with large amounts of data, due to the ability to perform one operation on several data simultaneously. The purpose of the study is to evaluate the effectiveness of using SIMD instructions to improve the performance of software code execution when processing large data sets compared to traditional software tools. The paper solves the following tasks: develop an algorithm for implementing the classical task of multiplying ultra-large (up to 36×106 bytes) square data matrices using the built-in Microsoft Visual Studio ISO/IEC C++20 library with SIMD technology to parallelise the program at the data level; study the performance of the developed algorithm when processing a significant amount of data compared to the traditional approach. By implementing a modified matrix multiplication algorithm using SIMD technology, it was possible to speed up the computation on a PC with an Intel Core i7-12700H processor by 4.8 times with a data volume of ~9×106 bytes. The obtained results will be taken into account in the development of application software, including for efficient computer models of technological processes and systems. [ABSTRACT FROM AUTHOR]
Copyright of Informatics & Mathematical Methods in Simulation / Informatika ta Matematičnì Metodi v Modelûvannì is the property of Odessa Polytechnic University and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Databáze: Complementary Index
Popis
Abstrakt:The current stage of development of processes and technologies requires continuous improvement of computer hardware performance, efficient use of its resources, processing of large amounts of data and support of the growing requirements of modern information systems. When processing large amounts of data, it is often necessary to use additional effective solutions to speed up information processing in addition to parallel computing. One such approach is to use the SIMD mechanism. The concept of SIMD instructions is a progressive solution for speeding up computations in tasks with large amounts of data, due to the ability to perform one operation on several data simultaneously. The purpose of the study is to evaluate the effectiveness of using SIMD instructions to improve the performance of software code execution when processing large data sets compared to traditional software tools. The paper solves the following tasks: develop an algorithm for implementing the classical task of multiplying ultra-large (up to 36×10<sup>6</sup> bytes) square data matrices using the built-in Microsoft Visual Studio ISO/IEC C++20 <immintrin.h> library with SIMD technology to parallelise the program at the data level; study the performance of the developed algorithm when processing a significant amount of data compared to the traditional approach. By implementing a modified matrix multiplication algorithm using SIMD technology, it was possible to speed up the computation on a PC with an Intel Core i7-12700H processor by 4.8 times with a data volume of ~9×10<sup>6</sup> bytes. The obtained results will be taken into account in the development of application software, including for efficient computer models of technological processes and systems. [ABSTRACT FROM AUTHOR]
ISSN:22235744
DOI:10.15276/imms.v13.no3-4.228