NVIDIA DGX SuperPOD
High-performance computing (HPC) cluster, specifically tailored to meet the demands of cutting-edge research.
NVIDIA DGX SuperPOD
The SMU NVIDIA DGX SuperPOD Advantage is a high-performance computing (HPC) cluster, specifically tailored to meet the demands of cutting-edge research.
This shared resource machine consists of 20 NVIDIA DGX A100 nodes, each with 8 advanced and powerful graphical processing units (GPUs) to accelerate calculations and train AI models. The O'Donnell Data Science and Research Computing Institute (ODSRCI) and Office of Information Technology (OIT) jointly manage and provide both access and support for this top of the line machine.
NVIDIA DGX SuperPOD Advantage Specifications
| Computational Ability | 1,644 TFLOPS |
| Number of Nodes | 20 |
| CPU Cores | 2,560 |
| GPU Accelerator Cores | 1,392,640 |
| Total Memory | 52.5 TB |
| Node Interconnect Bandwidth | 10 - 200 Gb/s Infiniband Connections Per Node |
| Work Storage | 768 TB (Shared) |
| Scratch Storage | 750 TB (Raw) |
| Archival Storage | N/A |
| Operating System | Ubuntu 20.04 |