Performance analysis of matrix calculation in parallel systems using OpenMP
Abstract
In view of the predominance of computational systems with multiple processing cores across the industry, this study aims to analyze some initial approaches, using the OpenMP API, for parallelizing the matrix multiplication. The results show that the strategy that considers the parallelization of the outer loop, compared to the intermediate one, reaches a better performance.
Keywords:
Performance measurements, evaluation and prediction, High-Performance Computing
References
A. Krause, G. Moro, L. S. (2016). Análise de desempenho da multiplicação de matrizes por strassen contra o método tradicional. In WSPPD 2016, 14th Workshop on Parallel and Distributed Processing.
Gove, D. (2011). Multicore Application Programming. Pearson Education, Inc.
M. Müller, B. Supinski, B. C. (2009). Evolving OpenMP in an Age of Extreme Parallelism. Springer.
Rathore, Y. and Kumar, D. (2014). Performance evaluation of matrix multiplication using openmp for single dual and multi-core machines. IOSR Journal of Engineering (IOSR- JEN), 4:56–59.
Gove, D. (2011). Multicore Application Programming. Pearson Education, Inc.
M. Müller, B. Supinski, B. C. (2009). Evolving OpenMP in an Age of Extreme Parallelism. Springer.
Rathore, Y. and Kumar, D. (2014). Performance evaluation of matrix multiplication using openmp for single dual and multi-core machines. IOSR Journal of Engineering (IOSR- JEN), 4:56–59.
Published
2021-05-06
How to Cite
LIBÓRIO, André; BALDASSIN, Alexandro.
Performance analysis of matrix calculation in parallel systems using OpenMP. In: REGIONAL SCHOOL OF HIGH PERFORMANCE COMPUTING FROM SÃO PAULO (ERAD-SP), 12. , 2021, Evento Online.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2021
.
p. 13-16.
DOI: https://doi.org/10.5753/eradsp.2021.16694.
