SuperServer SRS-48UAC-RTXPRO

AI Factory Solutions with NVIDIA RTX PRO 6000 Blackwell Server Edition

Key Applications

Artificial Intelligence (AI)
Design & Visualization
Deep Learning Inferencing
GPU Virtualization (VDI)
LLMs & Multi-modal LLMs

Key Features

Full-stack solutions based on reference architectures that include Supermicro systems, NVIDIA GPUs, NVIDIA software, and NVIDIA networking
Up to 256x NVIDIA RTX PRO™ 6000 Blackwell Server Edition GPUs providing up to 24TB GDDR7 memory
Compatibility with the NVIDIA software stack (NVIDIA AI Enterprise, NVIDIA Omniverse™, and NVIDIA Run:ai)
Plug-and-play solution with systems fully integrated into racks and tested before shipment and on-site deployment
NVIDIA Spectrum™-X Ethernet Compute Fabric, Converged Network, and Out-of-Band Management included
Supermicro AI Factory solutions are endorsed by NVIDIA for Infrastructure Configuration, Spectrum-X networking, and Software Reference Stack and based on the NVIDIA Enterprise Reference Architecture for RTX PRO 6000 Blackwell Server Edition.

...

...

Cluster Overview
Compute Nodes	4, 8, or 32 nodes (32–256 GPUs) per scalable unit
Compute Racks	Up to 4x compute racks per scalable unit
Networking Racks	Up to 1x networking rack per scalable unit
Rack Form Factor	48U 19"-width rack enclosure
Rack Dimensions	750mm x 1200mm
Total GPUs	Up to 256x NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs
Total GPU Memory	Up to 24TB GDDR7

Compute Node Configuration
System	SYS-422GL-NR
GPU	8x NVIDIA RTX PRO 6000 Blackwell Server Edition GPU
CPU	2x Intel® Xeon® 6972P Processor 96-Core 2.40GHz 480MB Cache (400W)
Memory	24x 64GB DDR5 6400MHz ECC RDIMM
Local Node Storage	Micron 7450 PRO 960GB NVMe
Networking	1x NVIDIA BlueField®-3 (B3220) and 4x NVIDIA BlueField-3 (B3140H)
Power	9.1kW (full load power draw)
System	SYS-522GA-NRT
GPU	8x NVIDIA RTX PRO 6000 Blackwell Server Edition GPU
CPU	2x Intel® Xeon® 6972P Processor 96-Core 2.40GHz 480MB Cache (400W)
Memory	24x 64GB DDR5 6400MHz ECC RDIMM
Local Node Storage	Micron 7450 PRO 960GB NVMe
Networking	1x NVIDIA BlueField®-3 (B3220) and 4x NVIDIA BlueField-3 (B3140H)
Power	9.1kW (full load power draw)
System	AS -5126GS-TNRT2
GPU	8x NVIDIA RTX PRO 6000 Blackwell Server Edition GPU
CPU	2x AMD EPYC™ 9575F Processor 64-Core 3.3GHz 256MB Cache (400W)
Memory	24x 64GB DDR5 6400MHz ECC RDIMM
Local Node Storage	Samsung PM9A3 960GB NVMe PCIe
Networking	1x NVIDIA BlueField®-3 (B3220) and 4x NVIDIA BlueField-3 (B3140H)
Power	8.6kW (full load power draw)

Compute Rack Configuration
Compute Nodes	4 or 8 Compute Nodes per rack
Estimated Rack Power	4 nodes per rack: ~36.6kW (SYS-522GA-TNR), ~33.3W(SYS-422GL-NR), and ~34.2kW (AS -5126GS-TNRT2). 8 nodes per rack: ~73.2kW (SYS-522GA-TNR), ~66.5W(SYS-422GL-NR), and 68.5kW (AS -5126GS-TNRT2).
Total GPUs	32 GPUs (4 nodes) or 64 GPUs (8 nodes) per rack
Total CPUs	8 CPUs (4 nodes) or 16 CPUs (8 nodes) per rack

Networking Overview
Compute Fabric	Up to 2x NVIDIA Spectrum-4 800GbE Ethernet Switch (SN5610)
Converged Network	Up to 2x NVIDIA Spectrum-4 800GbE Ethernet Switch (SN5610)
Out-of-Band Management	Up to 4x NVIDIA Spectrum 1GbE Ethernet Switch (SN2201)

Software Overview
Infrastructure Software	NVIDIA Run:ai
Application Software	NVIDIA AI Enterprise NVIDIA Omniverse™

Services
Onsite Deployment	Plug-and-play on-site deployment and validation
Rack Integration	Full rack-level design, component installation, and cabling System-level testing of each node (L11) Cluster-level multi-node testing (L12)

SuperCluster SRS-48UAC-RTXPRO New