Rice University researchers engaged in groundbreaking data-intensive science and engineering increasingly depend on access to real-time data analysis facilities required for their research. These research activities include image processing, computer vision, and machine learning, spanning multiple fields, such as geological sciences, statistics, computer science, and physics. Each of these problems areas or use cases can be addressed by shared computational infrastructure leveraging GPU accelerators for interactive computing. The system provides a significant resource for enabling science but also for educating the next generation of computational scientists in the latest GPU-computing techniques through the outreach of the Center for Research Computing.
The resource includes nine compute nodes, each with 40 cores, 384GB RAM, 4TB NVMe storage, and 8 NVIDIA Quadro RTX 6000 GPUs. The systems are interconnected via high-performance networking and hosted on a Science DMZ integrating them with the Open Science Grid as well as commercial cloud allowing both increased utilization as part of national OSG efforts and the ability to utilize cloud resources for load bursting. The system leverages an open-source software stack designed to support containerization, enabling each researcher to utilize their own unique set of software and toolkits while sharing common hardware and a common cloud access platform. Moreover, the infrastructure is part of a larger technology ecosystem that leverages federated identity and access management as part of InCommon, advanced networking with science DMZ, and Information Security Office that supports not only university data and technology security but includes targeted outreach for research data and protocol security.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
