A new high-performance computing (HPC) system that can do in seconds what would take your laptop hours or days is now online at Idaho State University.
Research Data Center staff recently activated Ragnarok, the facility’s latest HPC system. Ragnarok has eight Nivida RTX 3090 graphics processing units. Graphics processing units are specialized computer processors used to accelerate graphics rendering. Comparing Ragnarok to the average mainstream gaming laptop, Ragnarok has eight times the graphics processing power and is designed for a level of graphics computing that mainstream PCs lack.
“Idaho State researchers can use Ragnarok for many graphics-intensive processes, such as processing bare ground or canopy images collected using Lidar, data collected from drone flights , etc,” said Kindra Blair, Idaho State’s research systems administrator. “Future research could include work with neural networks and other machine learning algorithms.”
Ragnarok is one of four HPC systems available to researchers in the state of Idaho. He joins Minerva and Thorshammer; all housed on the Pocatello campus. Dustin McNulty, professor and chair of the physics department, used Minerva for particle physics and optics simulations. Using data from the simulations helped McNulty and his students develop new particle detection systems. The particle detectors have been used in nuclear physics experiments at ISU’s own Idaho Accelerator Center and around the world at the Thomas Jefferson National Accelerator Facility in Virginia, the Stanford Linear Accelerator Center in California, and the Mainzer Microtron facility. in Germany.
“I use high performance computing resources in almost every aspect of my research,” McNulty said. “Whether we’re designing a unique sensing system or trying to understand a physical or natural process, we set up and run detailed HPC simulations. Simulations give us a wealth of information, give more accurate predictions than traditional calculations, and really speed up our entire research process.
Meanwhile, Kathryn Turner, an assistant professor of biological sciences, and her students are using Thorshammer to help them analyze how populations of a plant species are related and how populations can spread across the landscape. Thorshammer is allowing them to develop a reference genome – a database of DNA sequences representative of the gene set of an idealized individual organism of a species – for the invasive weed, blue mustard. In addition to invasive weeds, Turner is also conducting similar studies on important native species such as greater sagebrush.
“Thorshammer and other high-performance computing resources are essential for the type of genomic analyzes we perform,” Turner said. “The datasets used in these analyzes are often hundreds of gigabytes in size and far, far too large to run on desktop computers. They would crash or take months to run. If we didn’t have Thorshammer, we would have had to find those kinds of resources elsewhere, which would most likely mean paying for it from scarce grant resources and competing for computer access.
The fourth, Falcon, located at the Idaho National Laboratory Collaborative Computing Center in Idaho Falls and connected to ISU via the Idaho Regional Optical Network. Falcon is billed as “one of the nation’s fastest university supercomputers” and “was ranked 97th on the world’s fastest supercomputers list when it was first rolled out in 2014.” according to a press release from the INL. Falcon was updated in 2017 and can perform a quadrillion – 1,000,000,000,000,000 – calculations per second.
“Falcon has been used by academic researchers since its inception, but the unique collaboration between Idaho’s three major research universities and focus on academic research allows many more researchers to use this system,” said said Michael Ennis, High Performance Computing Solutions Architect at State of Idaho. “Combining the graphics capabilities of Ragnarok with the extended computing power of Falcon and our other high-performance computers gives a huge boost to ISU’s research computing capabilities.”
For more details, visit the Idaho State University Research Data Center website.