The complete data center liquid cooling solution uses the latest high-density GPU servers and is equipped with high-performance CPUs and GPUs to accelerate the construction of AI factories
SAN JOSE, Calif. and HAMBURG, Germany, May 17, 2024 /PRNewswire/ --
Supermicro, Inc. (NASDAQ: SMCI ）As a full-service IT solutions manufacturer in AI, cloud, storage, and 5G/edge, we are committed to meeting our customers' demanding needs, including scaling AI and high-performance computing performance while reducing data center power consumption. Supermicro offers a complete solution for liquid cooling, including heat sinks, Cooling Distribution Units (CDUs), Cooling Distribution Manifold (CDMs), and entire cooling towers. Data center liquid-cooled servers and infrastructure enable rapid and significant reductions in data center power usage effectiveness (PUE), reducing overall power consumption by up to 40 percent.
"Supermicro will continue to work with our AI and HPC customers to bring the latest technologies, including a full range of liquid cooling solutions, to their data centers," said Charles Liang, President and CEO of Supermicro. Our complete liquid cooling solutions address cooling needs of up to 100kW per rack, reducing the total cost of ownership (TCO) of data centers and enabling higher density AI and high-performance computing. We bring the latest GPUs and accelerators to market with our Building Block Architecture, and we continue to bring new cabinet-level solutions to market with our trusted vendors, while shipping to customers with shorter lead times. "
Supermicro's application-optimized, high-performance servers are designed to house powerful CPUs and GPUs for simulation, data analytics, and machine learning. Supermicro 4U 8-GPU liquid-cooled servers deliver exceptional performance, delivering petaflop-level AI computing power in a high-density form factor with NVIDIA H100/H200 HGX GPUs. Supermicro will soon be offering liquid-cooled Supermicro X14 SuperBlades, Rack X14 Hyper and Supermicro X14 BigTwin in 8U and 6U configurations. Several HPC-optimized server platforms will be available in a compact, multi-node form factor and support for the Intel Xeon 6900 with P-cores.
Learn more about how Supermicro Rack-Level Consolidation Services can help you reduce costs and optimize your data center:
In addition, Supermicro continues to lead the industry with a broad portfolio of liquid-cooled MGX products. Supermicro has also confirmed support and availability of Intel's latest accelerators, including the new Intel® Gaudi® 3 accelerator, as well as AMD's MI300X accelerator. With up to 120 nodes per cabinet, Supermicro SuperBlades® can run large-scale high-performance computing applications in just a few cabinets. Supermicro will be showcasing a wide range of server types, including the Supermicro X14 system with Intel® Xeon® 6 processors.
Supermicro will also be exhibiting at ISC 2024 and demonstrating a variety of solutions designed for high-performance computing and AI environments. The new 4U 8-GPU liquid-cooled servers with NVIDIA HGX H100 and H200 GPUs are the highlight of Supermicro's product line. These and other servers will also support NVIDIA B200 HGX GPUs. The new system with high-end GPUs uses high-speed HBM3 memory to bring data closer to the GPU than previous generations, accelerating AI training and high-performance computing simulations. With the superior density of 4U liquid-cooled servers, a single cabinet can provide: 8 servers x 8 GPUs x 1979 Tflop FP16 (with sparsity technology) = more than 126 petaflops of computing power. The Supermicro SYS-421GE-TNHR2-LCC is powered by two 4th or 5th Gen Intel Xeon processors, while the AS-4125GS-TNHR2-LCC is powered by two 4th Gen AMD EPYC™ CPUs.
The new AS-8125GS-TNMR2 server provides users with access to eight AMD Instinct™ MI300X accelerators. The system is also powered by two AMD EPYC™ 9004 series processors with up to 128 cores/256 threads and up to 6TB of memory. Within each AMD Instinct MI300X accelerator, each GPU has 192GB of HBM3 memory, all connected via an AMD Universal Baseboard (UBB 2.0). In addition, the new AS-2145GH-TNMR-LCC and AS-4145GH-TNMR APU servers are designed to accelerate high-performance computing workloads with the MI300A APU. Each APU combines high-performance AMD CPUs, GPUs, and HBM3 memory in a single system, providing 912 AMD CDNA™ 3 GPU compute units, 96 "Zen 4" cores, and 512GB of unified-architecture HBM3 memory.
At ISC 2024, Supermicro 3U servers with Intel Gaudi 8 AI accelerators will be unveiled. The new system is designed for AI training and inference, and can be directly connected to traditional Ethernet Fabric fabrics. Each Intel Gaudi 3 accelerator incorporates 24 200 Gb Ethernet ports for flexible, open-standard networking and built-in 128GB of HBM2e high-speed memory. The Intel Gaudi 3 accelerator efficiently scales from a single node to thousands of nodes vertically and horizontally to meet the broad needs of GenAI models. Supermicro's petabyte-scale storage systems are critical to large-scale high-performance computing and AI workloads, and these systems will also be on display at the conference.
In terms of data center management software, Supermicro will also be showcasing Supermicro SuperCloud Composer and demonstrating how to monitor and manage the status of the entire data center, including all liquid-cooled servers, from a single console.
For more information on Supermicro's participation at ISC 2024, please visit:
concerning Super Micro Computer, Inc.Supermicro (NASDAQ: SMCI) is a global leader in application-optimized comprehensive IT solutions. Headquartered and operated in San Jose, California, Supermicro delivers market-leading innovations for enterprise, cloud, AI, and 5G telco/edge IT infrastructures. We are a full-service IT solutions manufacturer, offering servers, AI, storage, IoT, switch systems, software and support services. Supermicro's motherboard, power supply, and chassis design expertise further optimizes our development and production to enable next-generation innovation from the cloud to the edge for our global customers. Our products are designed and manufactured by in-house teams (in the U.S., Asia and the Netherlands) to reduce total cost of ownership (TCO) through product design optimization, reduce environmental impact through green computing technology, and achieve excellent manufacturing scale and efficiency in global operations. The award-winning Server Building Block Solutions® portfolio enables customers to choose the right model from the most diverse system portfolio to get the best performance for their workloads and applications. The multi-system product line is made up of highly resilient, reusable building blocks that support a wide range of hardware form factors, processors, memory, GPUs, storage, networking, power consumption, and cooling solutions (air conditioning, free air cooling, or liquid cooling).
Supermicro、Server Building Block Solutions和We Keep IT Green皆为Super Micro Computer, Inc. 的商标和/或注册商标。
All other brands, names and trademarks are the property of their respective owners.