A parallel algorithm for matrix fast exponentiation based on MPI

Autor:	Ji Zhang, Xu-chuan Zhou
Rok vydání:	2018
Předmět:	Matrix (mathematics) Speedup 010201 computation theory & mathematics Computer science Server Message passing Message Passing Interface Parallel algorithm 0102 computer and information sciences Parallel computing 01 natural sciences Exponentiation by squaring Matrix decomposition
Zdroj:	2018 IEEE 3rd International Conference on Big Data Analysis (ICBDA).
DOI:	10.1109/icbda.2018.8367669
Popis:	Matrix Fast Exponentiation Algorithm (MFEA) is widely used as an algorithm for fast solving the matrix power. But its efficiency can not meet the needs of the big data application. In this paper, we analyze the parallel potential of MFEA and propose a Parallel Matrix Fast Exponentiation Algorithm (PMFEA) based on message passing interface (MPI). PMFEA uses data partitioning to improve the processing capacity for the big data. The experimental results show that PMFEA achieves about 8.41X speedup compared to MFEA on 32-core server.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::3fa751eba8d630cf0ab877c2eb427adb https://doi.org/10.1109/icbda.2018.8367669 Zobrazit plný text záznamu