Comparison of Time Variability in Different Matrix Multiplication Algorithms using Parallelization and Divide and Conquer

• Victor H. de Oliveira UTFPR
• Rogério Aparecido Gonçalves UTFPR
• João Fabrício Filho UTFPR

DOI: https://doi.org/10.5753/eradsp.2024.239870

Abstract

Matrix multiplication is a common application in the computational field. The objective of this work is to compare distinct approaches implementing matrix multiplication in relation to time execution. Dynamic and static allocation techniques, blocking and parallelization with multiple threads are covered. The strategy involving blocking implemented together with parallelization achieved performance gains 75.12% greater than the sequential matrix multiplication approach and 4.5% faster than parallel execution without the use of blocking.

References

Huss-Lederman, S., Jacobson, E. M., Tsao, A., Turnbull, T., and Johnson, J. R. (1996). Implementation of strassen’s algorithm for matrix multiplication. In Supercomputing, USA. IEEE Computer Society.

Pacheco, P. S. (2011). An Introduction to Parallel Programming. Morgan Kaufmann Publishers.

Supriya P. Mali, Sonali Dohe, P. R. (2019). Memory management techniques: Static and dynamic memory allocation. International Journal of Current Engineering and Technology, 9(1):92–94.