The DGX On-Prem AI Supercomputer runs intensive machine learning and deep learning workloads. It consists of on-premise NVIDIA DGX systems (sometimes referred to as the DGX On-Prem system) to run these workloads in a highly scalable fashion. By requesting access to the system, users will be able to run Jupyter notebooks, JupyterLab, as well as Python scripts that are particularly GPU-intensive. Additionally, any of the available containers that are provided by NVidia on the DGC Cloud can be pulled and used with the system.
Features
...
24 NVIDIA DGX A100 systems (each with 8 NVIDIA A100 GPUs)
3 Lambda scalers (each with 8 NVIDIA L40s)
Over 6 petabytes of NetApp storage
Available to
...
FacultyUAlbany Researchers
Cost
...
This resource is available to all faculty free of charge. In the future, a paid prioritization option will be made available as well.
More Information
...
Get started: Request access by completing a DGX On-Prem Computation Request
Learn more by visiting the DGX On-Prem How-To page
For more information on DGX On-PremAI Plus Supercomputing Cluster, contact the ITS Service Desk