Super Micro Computer, Inc. expands manufacturing for NVIDIA Vera Rubin platforms
Super Micro Computer, Inc. expanded its manufacturing capacity and liquid-cooling capabilities in collaboration with NVIDIA to speed deployment of liquid-cooled Artificial Intelligence (AI) infrastructure.
The company said its Data Center Building Block Solutions (DCBBS) approach streamlined production, offered customization options, and shortened time-to-deployment for data center-scale systems designed around the NVIDIA Vera Rubin and Rubin platforms.
The PR described two flagship systems: the NVL72 SuperCluster, which unifies 72 NVIDIA Rubin GPUs and 36 NVIDIA Vera CPUs with NVIDIA ConnectX-9 SuperNICs and NVIDIA BlueField-4 DPUs and uses NVLink 6 and NVIDIA Quantum-X800 InfiniBand or Spectrum-X Ethernet, and a 2U liquid-cooled HGX Rubin NVL8 system that houses eight GPUs and supports high-density 2U busbar designs with advanced Direct Liquid Cooling (DLC) technology.
The release listed platform features such as NVIDIA Vera Central Processing Unit (CPU) specs (88 cores/176 threads, 1.2 TB/s LPDDR5X memory bandwidth, 1.8 TB/s NVLink-C2C), a 3rd-generation Transformer Engine, 3rd-generation confidential computing, a 2nd-generation Rack Alignment System (RAS) engine, Spectrum-X Ethernet Photonics models SN6800, SN6810, and SN6600, and Supermicro-based Petascale all-flash storage and JBOF solutions supporting BlueField-4 DPUs.
“Supermicro's long-standing partnership with NVIDIA and our agile building block solutions enable us to bring the most advanced AI platforms to market faster than others,” said Charles Liang, president and CEO of Supermicro. “With expanded manufacturing and industry-leading liquid-cooling expertise, we're empowering hyperscalers and enterprises to deploy the NVIDIA Vera Rubin and Rubin platforms infrastructure at scale with unmatched speed, efficiency, and reliability.”
The companies described plans to use expanded manufacturing capacity and the end-to-end liquid-cooling technology stack together with the modular DCBBS architecture to streamline configuration, validation, and deployment of fully liquid-cooled NVIDIA Vera Rubin and Rubin platforms.