Feature - Midwest Grid Workshop |
||||
|
|
||||
Adjacent to the skyscrapers that make up the Chicago skyline and overlooking the "Little Italy" neighborhood, about 70 participants convened at the University of Illinois at Chicago March 24 and 25 for the Midwest Grid Workshop. Organized by Open Science Grid, TeraGrid and their partners, the attendees ranged from undergraduates with a passing knowledge of grid computing to professionals who already use a grid for science applications and are interested in using it in more powerful ways. Professor Muhammad Ali of Tuskegee University in rural Alabama, United States, came to the school with three goals: to learn about grid computing in general; to bring back information for science applications at the university that could benefit from grid technology; and to find out if students in his parallel processing course could run their lab exercises on the grid, so that Tuskegee wouldn't need to maintain a local cluster. He said he now knows that his goals were reasonable, but further collaboration is needed to achieve some of them. "This workshop met and exceeded my expectations for getting practical training," says Ali. "Tuskegee has several departments that are interested in processing large data sets, but are handicapped due to lack of resources. I think it would be very beneficial to hold a workshop at Tuskegee for our faculty and students."
In addition, Ali was excited to learn about virtual machines that are being developed and might be usable by his students for their lab exercises. "We would be very happy to participate during trials for these machines and give our feedback." Ali and his fellow workshop participants heard lectures on national grid infrastructures, workflow systems, data mining and other topics. They also participated in hands-on computer labs. In one exercise, students submitted grid and Condor DAGMan jobs via Globus to a local grid, then submitted jobs over OSG to machines at the San Diego Supercomputing Center. Uri Hasson of the Human Neuroscience Laboratory at the University of Chicago, United States, is a more experienced grid user who has been running a cognitive neuroscience application. This application processes a series of measurements, attempts to rule out chance readings, and determines the reliability of differences between activity level readings in particular areas of the brain. He was interested in finding out about OSG's capabilities and the direction in which its development is headed. "As a person on the science side of things," Hasson says, "this grid school provided me with a more systematic introduction to the capabilities of the OSG and to what goes on 'under the hood.' I now see that it can greatly assist me in my research. If I end up heading a research lab, I plan to do what I can to get it on the OSG with a fast connection." - Marcia Teckenbrock, OSG This article originally appeared as an OSG News item. |
