AI HardwarePDFPublished
GPU Cluster Reference Design
Rack layout, power, cooling and fabric design for a 64-accelerator training pod.
About this resource
Reference architecture covering rack elevation, 415V power distribution, rear-door heat exchangers, network fabric topology, cabling schedules and a commissioning test plan. Source: https://docs.nvidia.com/nvidia-dgx-superpod-data-center-design-dgx-h100.pdf
datacenterGPUinfrastructure