DGX SuperPOD Architecture

The DGX SuperPOD architecture is a combination of DGX systems, InfiniBand and Ethernet networking, management nodes, and storage. Figure 1 shows the rack layout of a single SU (Scalable Unit). Each SU requires a Thermal Design Power (TDP) of 1.2 Megawatts (MW). Generally, the data center should meet or exceed Uptime Institute Tier 3 design standards, or alternatively the TIA942-B Rated 3 or EN50600 Availability Class 3 design standards, including concurrent maintainability and no single point of failure.

Each DGX GB200 leverages hybrid cooling with both direct liquid cooling and air cooling to manage the amount of heat produced by each rack. Given DGX SuperPOD with DGX GB20 is a turnkey AI data center scale product, NVIDIA has provided datacenter design guidance and fully integrated operational technology (OT) integration with NVIDIA Mission Control software stack.

Figure 1 Complete Rack Layout in One Row

This reference architecture is focused on the design of a single SU (8 DGX GB200 rack systems). DGX SuperPOD can scale to much larger configurations up to and beyond 128 racks with 9216 GPUs.

Contact your NVIDIA representative for information regarding DGX SuperPOD solutions of four SUs or more.