The Prototype National Research Platform (PNRP) is a Category II NSF-funded resource integrated into the Nautilus cluster operated jointly by the San Diego Supercomputer Center at UC San Diego, the Massachusetts Green High Performance Computing Center (MGHPCC) and the University of Nebraska–Lincoln (UNL). The system features a variety of hardware, including FPGAs, GPUs, and NVMe storage. Each of the three sites (SDSC, UNL, and MGHPCC) includes ~1 PB of usable disk space. The three storage systems function as data origins of the CDN, providing data access anywhere in the country within a round-trip delay of ~10ms via use of network caches at three sites and five Internet2 network colocation facilities.
The Nautilus cluster serves the broader National Research Platform (NRP) which is a community-owned research and education platform connecting researchers and educators to foster collaboration, accelerate innovation, and share resources.
The PNRP contribution to the cluster comprises of 1) a HPC subsystem at SDSC with 8 HGX A100 servers with 8 80G A100 GPUs, 512G memory, and 1TB NVMe storage per server; 32 Alveo U55C FPGAs available composed nodes via GigaIO fabric; 122TB of FabreX connected NVMe; 2) two FP32 subsystems, one each at UNL and MGHPCC, each with 18 GPU nodes with 8 A10 GPUs, 512G memory, and 8TB NVMe per node; and 3) 8 distributed data caches of 50TB each. In addition, the distributed Kubernetes cluster architecture enables other institutions to incorporate their own resources into the cluster.
Members get updates about announcements, events, and outages on the PNRP ACCESS resource.
Associated Resources
-
PNRP
The Prototype National Research Platform (PNRP) is a Category II NSF-funded system integrated into the Nautilus cluster operated jointly by the San Diego Supercomputer Center at UC San Diego, the Massachusetts Green High Performance Computing Center (MGHPCC) and the University of Nebraska–Lincoln (UNL). The system features a novel, extremely low-latency fabric from GigaIO that allows dynamic composition of hardware, including FPGAs, GPUs, and NVMe storage. Each of the three sites (SDSC, UNL, and MGHPCC) includes ~1 PB of usable disk space. The three storage systems function as data origins of the CDN, providing data access anywhere in the country within a round-trip delay of ~10ms via use of network caches at three sites and five Internet2 network colocation facilities. The Nautilus cluster serves the broader National Research Platform (NRP) which is a community-owned research and education platform connecting researchers and educators to foster collaboration, accelerate innovation, and share resources. The PNRP contribution to the cluster comprises of 1) a HPC subsystem at SDSC with 8 HGX A100 servers with 8 80G A100 GPUs, 512G memory, and 1TB NVMe storage per server; 32 Alveo U55C FPGAs available composed nodes via GigaIO fabric; 122TB of FabreX connected NVMe; 2) two FP32 subsystems, one each at UNL and MGHPCC, each with 18 GPU nodes with 8 A10 GPUs, 512G memory, and 8TB NVMe per node; and 3) 8 distributed data caches of 50TB each. In addition, the distributed Kubernetes cluster architecture enables other institutions to incorporate their own resources into the cluster.
Members get updates about announcements, events, and outages.
