Cerebras’s Wafer Scale Engine, with a baseball for size comparison.

Access and AI research begins on the new PSC Neocortex system, an advanced AI computing system for science and engineering that leverages Cerebras Systems’ revolutionary Wafer Scale Engine

The Pittsburgh Supercomputing Center (PSC) has deployed Neocortex, a unique high performance artificial intelligence (AI) system designed to revolutionize scientific AI research. A number of cutting-edge science and engineering projects are now running on Neocortex as a part of the system’s early user program, which began in mid-February. These projects include efforts in drug discovery, genomics, molecular dynamics, climate research, computational fluid dynamics, signal processing and medical imaging analysis.

“At PSC, we are extremely pleased to reach this milestone and enable early user access to Neocortex,” said project leader Paola Buitrago, PSC’s Artificial Intelligence and Big Data Director. “Deploying this outstanding and innovative system would not have been possible without a formidably strong collaboration between users, vendors and our outstanding PSC team for which we are extremely grateful.”

Funded by a $5-million grant from the National Science Foundation in June 2020, Neocortex introduces fundamentally new hardware to greatly speed up AI research. PSC, a joint research organization of Carnegie Mellon University and the University of Pittsburgh, built the new system in partnership with Cerebras Systems and Hewlett Packard Enterprise (HPE). Neocortex is available at no cost to selected researchers advancing open research.

“The Neocortex program introduces a spectacular cutting-edge supercomputer resource to accelerate our research to combat breast cancer,” said Dr. Shandong Wu, Associate Professor of Radiology and Director of the Intelligent Computing for Clinical Imaging (ICCI) Lab at the University of Pittsburgh, who leads a team employing deep learning on Neocortex for high-resolution medical imaging analysis for breast cancer risk prediction. “The program also has a very responsive support team to help us resolve issues and has enabled great progress of our work.”

Neocortex goes beyond the technologies that have powered much of the advancement in AI since 2012. The system does this by exploring a revolutionary combination of Cerebras’ CS-1 systems and Wafer Scale Engine (WSE) processors, which are designed specifically to accelerate AI, and a large-memory HPE Superdome Flex server for data handling capability. Cerebras’ WSE processors are the largest computer chips ever built and power the Cerebras CS-1 system. Depending on workload, the CS-1 delivers hundreds or thousands of times more performance than legacy alternatives, and it does so at a fraction of the power draw and space.

“As early users, we are working to use Neocortex to train novel graph neural algorithms for very large graphs, such as those found in social networks,” said Dr. John Wohlbier of the Emerging Technology Center at the CMU Software Engineering Institute, another early user. “We are excited to explore to what extent will AI-specific ASICs, such as the one in Neocortex, enable model training and parameter studies on a much faster timescale than resources found in a typical datacenter or the cloud.”

The balanced Neocortex system democratizes access for researchers nationwide to game-changing compute power for training, the most time-consuming step in deep learning research, to be much faster, even interactive. The innovative architecture has already inspired similar systems across the globe.

“We are tremendously excited to be a part of the Neocortex program and welcome such a wide range of innovative and ambitious AI projects on to our CS-1 systems,” said Andy Hock, VP Product at Cerebras Systems. “We’ve also been thrilled to see strong engagement from the Neocortex user community – this engagement and feedback from this user community will be invaluable for our future planning and work together to accelerate AI for science even further.”

The system’s unique AI capabilities for advanced research and development, and its speed, will remove barriers to AI innovation, supporting the Executive Order on Maintaining American Leadership in Artificial Intelligence and the American Artificial Intelligence Initiative.

To learn more about Neocortex, please email neocortex@psc.edu, visit the project website, and/or subscribe to the project newsletter.