Performance Evaluation of Multiple Approximate String Matching Algorithms Implemented with MPI Paradigm in an Experimental Cluster Environment

Author

Michailidis, Panagiotis D. ; Margaritis, Konstantinos G.

Author_Institution

Dept. of Appl. Inf., Univ. of Macedonia, Thessaloniki

fYear

2008

fDate

28-30 Aug. 2008

Firstpage

168

Lastpage

172

Abstract

This paper describes and analyzes two high performance multiple string searching algorithms that we have optimized for cluster of heterogeneous workstations. The proposed parallel algorithms are based on dynamic master-worker programming paradigm. Experimental results show that these parallel algorithms are efficient: their search time is cut down significantly, whereas their communication cost stay at a low level.

Keywords

application program interfaces; message passing; parallel algorithms; string matching; MPI paradigm; approximate string matching algorithms; cluster environment; communication cost; dynamic master-worker programming paradigm; heterogeneous workstations; multiple string searching algorithms; performance evaluation; Clustering algorithms; Dictionaries; Dynamic programming; Field programmable gate arrays; Informatics; Information filtering; Load management; Parallel algorithms; Parallel programming; Workstations; MPI; approximate string matching; cluster of heterogeneous workstations;

fLanguage

English

Publisher

ieee

Conference_Titel

Informatics, 2008. PCI '08. Panhellenic Conference on

Conference_Location

Samos

Print_ISBN

978-0-7695-3323-0

Type

conf

DOI

10.1109/PCI.2008.13

Filename

4621557