Evaluating BLAST runtime using NAS-based high performance clusters

Sadiq M. Sait*, M. Al-Mulhem, Raed Al-Shaikh

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

The Basic Local Alignment Search (BLAST) is one of the most widely used bioinformatics programs for searching all available sequence databases for similarities between a protein or DNA query and predefined sequences, using sequence alignment technique. Recently, many attempts have been made to make the algorithm practical to run against the publicly available genome databases on large parallel clusters. This paper presents our experience in evaluating both the serial and parallel BLAST algorithms onto a large Infiniband-based diskless High Performance Cluster (HPC) that offers lower hardware cost and improved reliability, as opposed to traditional disk full clusters. The paper also presents the evaluation methodology along with the experimental results to illustrate the scalability of the BLAST algorithm on our HPC system. For our measurement and comparison, we considered cluster sizes up to 32 compute nodes. Our results show that BLAST runtime can still be retained with the use of the diskless clusters, while improving the runtime reliability.

Original languageEnglish
Title of host publicationProceedings - CIMSim 2011
Subtitle of host publication3rd International Conference on Computational Intelligence, Modelling and Simulation
Pages51-56
Number of pages6
DOIs
StatePublished - 2011

Publication series

NameProceedings - CIMSim 2011: 3rd International Conference on Computational Intelligence, Modelling and Simulation

Keywords

  • BLAST
  • Diskless clusters
  • Experimental Performance
  • Genomes
  • HPC Reliability
  • Infiniband

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computational Theory and Mathematics
  • Modeling and Simulation

Fingerprint

Dive into the research topics of 'Evaluating BLAST runtime using NAS-based high performance clusters'. Together they form a unique fingerprint.

Cite this