William Burke's Web page
Email Address: wburke999@msn.com

Group Project: Building, Programming and Administering Beowulf Clusters
Individual Project: Cluster Computing Applications Project Parallelizing BLAST
![]()
| Undergraduate:
York College, C.U.N.Y. |
Major:
Information Systems Management |
Mentors Name: John Mugler, Stephen Scott |
Mentor's Email: |
Upon the end of my junior year, I was chosen along with 2 other students (Lori Collins and Amara Diggs) from York College, C.U.N.Y. for the Research Alliance for Minorities (RAM) program to intern at Department of Energy's Oak Ridge National Laboratory in the Computer Science and Mathematics division. Our mentors Stephen Scott, PH.D. and his team of researchers (John Mugler, Thomas Naughton, and Brian Luethke) trained us on the latest cluster computing software and parallel processing technologies. We spent the first part of our summer learning cluster infrastructure utilizing Open Source Cluster Application Resources (OSCAR). This was vital since the rest of the project consisted of utilizing clusters for various problems. The first part of the project included learning RedHat Linux 7.2, learning OSCAR and its included toolsets, testing OSCAR 1.3 beta 3 version, logging bugs found during the OSCAR installation process, communicating these bugs to the OSCAR group, and updating the installation document. The second half of our summer was spent building and administering clusters. We then began to apply our understanding of OSCAR and clusters to start individually working on separate projects. I began investigating a Bioinformatics application: Basic Local Alignment Sequence Tool (BLAST) used for genetic sequencing. Bioinformatics is an emerging discipline, which combines computer science computational science and various aspects of biological science. The staple of bioinformatics is the creation and maintenance of databases of biological information. I downloaded different versions of BLAST and ran it to my personal cluster and the eXtreme Tennessee Oak Ridge Cluster (XTORC), a 65 (Pentium IV) nodes Linux based Beowulf cluster (a supercomputer), capable of 129.7 GFLOPS. My desire to further my education has really increased as a result of this experience and my goal is to continue my education until I have attained a Ph.D. I have the intentions of mentoring high school kids as well as other undergraduates to introduce them to the wonderful world of research so that they too will have the same desire to further their education.
I learned the latest technology in Parallel Computing and Beowulf Clustering. My hands-on training lasted for 11 weeks which included:
Learning Cluster Infrastructure, Parallel Processing, and Cluster Applications.
Installing the Open Source Cluster Applications Resource (OSCAR) environment.
Building, Installing, Administering, and Monitoring RedHat 7.2 Linux base Beowulf Clusters.
Using the necessary components for High Performance Cluster (HPC) computing.
Researching the BLAST algorithm, a heuristic algorithm used in the Bioinformatics industry for matching query text strings to a database of text strings, to investigate the feasibility of parallelizing its algorithms for faster text string searches and analysis.
Please feel free to view my PowerPoint presentation and a poster, a research paper and this web page.




