The NIU Center for Research Computing and Data (CRCD) is commissioning a new $1.2 million, high-performance computing facility (HPC), known as METIS.
“We want to provide our faculty with the sophisticated tools they need to stay on the cutting-edge of research and scholarship in their respective fields,” said Gerald C. Blazey, NIU vice president of Research and Innovation Partnerships. “Relative to its predecessor, METIS increases NIU HPC capacity by an order of magnitude.”
Professor Bela Erdelyi (in the blue shirt), director of the CRCD, and Sergey Uzunyan, the CRCD’s director of science and engineering.
For the techies among us, METIS will increase CRCD’s CPU capacity from 1.2 teraflops to 16.4 teraflops and GPU capacity from 21.4 teraflops to 292 teraflops. With a capability of 3 X 1014 mathematical calculations per second and an ultrafast communications network, it will be able to handle enormous data sets and complex computational investigations.
The METIS computing facility, which replaces the 10-year-old Gaea hybrid compute cluster, was purchased with university funds.
“To invest in such a facility makes sense for our research community,” said NIU Physics Professor Bela Erdelyi, director of the CRCD, which aims to serve as a team of experts to broadly support the diverse research computing and data needs on campus.
“These high-performance systems allow people to do research they otherwise couldn’t or get results back much faster, and that’s a win for everybody,” Erdelyi said. “The previous system was a decade old, and the hardware evolves so fast that a new one was needed. NIU scientists are now tackling many problems that go beyond the capacity of Gaea.”
Erdelyi expects METIS to start onboarding users later this month. Use of the facility is free of charge to NIU faculty.
In the past, the most common users of the CRCD have come from the STEM fields. Gaea supported nearly 100 scholarly and research projects, including weather and climate modeling, advanced accelerator topics, analysis of high-content biological images, simulations of charged particle beam dynamics, and application of computational materials science in materials design and physical/mechanical behavior of nanomaterials and nanostructures.
Erdelyi and Senior Research Scientist Sergey Uzunyan, the CRCD’s director of science and engineering, hope to expand use of METIS into more fields, ranging from the humanities to economics.
High-performance computing could be used, for example, for artificial-intelligence projects involving music, painting and writing; or even to build models of human consciousness. The new cluster of modern processors can perform large-scale financial forecasting or GPU-hungry art projects like visualizations of historical battlefields and back-in-time map travels.
“With the rapid advancement of artificial intelligence and machine-learning technologies, the demand for high-performance computing resources has never been greater,” Erdelyi said. “CRCD resources can help to significantly reduce the time and costs associated with running complex and resource-intensive simulations, modeling and data-analysis tasks. With the ability to process large amounts of data quickly and efficiently, we can accelerate our research, publish high-impact papers and attract top talent to our institution.”
“The system will benefit projects requiring fast parallel analysis of large data volumes,” Uzunyan adds. “Along with traditional STEM projects, the system can help conduct research on speech and visual recognition, medicine, psychology, finance, climate, environmental problems and insurance.”
Both Erdelyi and Uzunyan are available to speak with faculty about leveraging METIS for their research.
“People might be uncertain on how they could use high-performance computing,” Erdelyi said. “My pitch to them would be to invite Sergey or me to meet with your research group, department or college so you can have better understanding of our new system and how you might be able to apply it to your research.”
Faculty interested in learning more can contact Bela Erdelyi at berdelyi@niu.edu or Sergey Uzunyan at suzunyan@niu.edu.
