Computer Science Department,
College of Arts and Sciences
Acquisition of an Instrument for Research in
Irregularly Parallel Big Data Computation
(October 11, 2013) New Mexico State University will soon see the arrival in the College of Arts and Sciences of an instrument for research in irregularly parallel big data computation, thanks to Jonathan Cook, associate professor of computer science. Cook was recognized for his work at an NMSU Research Rally Friday, Oct. 11. The project will be funded by a National Science Foundation Major Research Instrumentation grant.
Cook’s $224,000 NSF grant will help acquire a computational instrument designed to support data-driven graph computations. “It’s for big data research,” he said. “My area of interest is software engineering: what’s it’s doing, how it’s behaving and how it can improve. I generate lots of data. The instrument has computation nodes that have large amounts of memory, with whole hosts of memory capacity at each node to handle that data and to allow the computation fast access to the data.” The type of analytical work Cook and his colleagues do is complimentary to traditional computer science studies. The device will also be used as a learning tool for students. “There is a tremendous need for students who understand data analysis, and they need access to computational resources and instruments like this,” he said. “We are already introducing a big data analysis graduate course this year. The instrument will provide a platform to deploy what we’ve learned into a bigger environment. This is really a research machine; it’s not a machine that’s big enough to, for example, process a large corporation’s amount of data.”
Cook’s research could potentially improve software and improve the performance of anything from scientific application software to video games. “Even improving the software performance of video games could have an economic impact,” he said. “We get sample applications that represent some type of software we want to analyze. Then we bring those in-house and collect data from them. I think the most fascinating part is the amount of data we could actually process. It’s hard to relate, because we throw out numbers like gigabyte or megabyte, and to really understand that number is really hard. The example I’ve used is that we generate 2.5 exabytes of data per day, worldwide. If you had one exa-millimeter, that would get you to Jupiter from the sun and back 500 times. It’s incredible to visualize that number.”
– Article by Isabel A. Rodriguez; photo by Darren Phillips. See more at newscenter.nmsu.edu.
Project funded through a major research instrumentation grant from the National Science Foundation.