New infrastructure at the University of Maryland is expanding high-performance computing access for undergraduate students while allowing machine learning researchers to model more data than ever before.
A computing cluster consisting of 330 Nvidia GPUs—graphical processing units that can perform multiple, simultaneous computations—has nearly doubled the infrastructure’s capacity. The interconnected GPUs can solve complex problems in computer vision, cybersecurity and more.
This added computational power has democratized access for undergraduate students, who previously could not use machine learning clusters reserved mainly for graduate students and faculty, said Derek Yarnell, director of computing facilities at the University of Maryland Institute for Advanced Computer Studies (UMIACS).
Yarnell, who led the installation, added that the new cluster will let machine learning researchers at any level tackle larger data sets with longer and more complicated algorithms.
“This is a significant increase across the board for our entire research community,” he said. “But the cluster’s biggest impact will be in providing new professors and students with access to these powerful tools.”
Mohammad Nayeem Teli, a senior lecturer in the Department of Computer Science, will employ the cluster in his “Introduction to Machine Learning” course this fall. The course requires students to work with large data sets that require more powerful computational resources as they become more complex. He said that the new cluster will enable students to run their models faster and enhance their overall performance.
A collaborative effort between UMIACS and the Department of Computer Science, the $2 million cutting-edge cluster was purchased with remaining computing infrastructure funds from the construction of the Brendan Iribe Center for Computer Science and Engineering.
“We’re excited about expanding our infrastructure to support more of our undergraduate students,” said Mihai Pop, director of UMIACS and a professor of computer science. “This enhancement will not only directly benefit our undergraduates but advance the level and reputation of machine learning research at the University of Maryland as a whole.”
Maryland Today is produced by the Office of Marketing and Communications for the University of Maryland community on weekdays during the academic year, except for university holidays.
Faculty, staff and students receive the daily Maryland Today e-newsletter. To be added to the subscription list, sign up here:Subscribe