Stony Brook University celebrated the opening of its astonishing new SMART Cluster, a dual use GPU Cluster for both machine learning and visualization, in a ribbon-cutting ceremony on campus November 4.
The high-tech cluster will facilitate powerful AI-enhanced computational capabilities, high-end visual computing and analytics, and acceleration of data-intensive and compute-intensive processes.
“It is a world-class cluster designed specifically for machine learning as well as visualization,” said Arie Kaufman, Distinguished Professor in Stony Brook’s Department of Computer Science and Chief Scientist of the Center of Excellence in Wireless and Information Technology (CEWIT).
“According to NVIDIA, our SMART cluster is the fastest and the largest among academic institutions.” he said. “It can execute three quadrillion operations per second.”
The new technology also drives the freshly upgraded Reality Deck 2.0, the largest immersive display in the world with 1.5 billion pixels, as well as Silo, a cylindrical immersive stereo display with 0.6 billion pixels.The SMART Cluster is also the first hardware-accelerated ray-tracing cluster for real-time cinematic quality rendering, allowing scientists, engineers and physicians to visualize huge amounts of data in a new way.
The system could be utilized for a powerful range of research endeavors, ranging from climate change simulations to law enforcement. The technology will also enable medical advances, such as virtual colonoscopies so detailed that the computer could potentially detect polyps without human intervention.
“This kind of cluster is a supercomputer on steroids,” Kaufman said. “It will excite not only the researchers on this campus, but also every scientist who is seeking to analyze their data, every engineer looking to verify their design, and every physician deciding on a diagnosis.”
