Super Micro Computer, Inc. reveals a range of servers to meet IT requirements for modern workloads.
The company’s Building Block Server methodology enables a first-to-market delivery with the technology from Intel, AMD, and NVIDIA. Purpose-built servers deliver performance for a range of AI, cloud, and 5G workloads, from the data center to the edge.
“As we expand our production capacity to meet the rapidly growing demand of high-performance large-scale AI infrastructure and cloud data centers, Supermicro delivers the industry’s most innovative and advanced systems integrated as a turn-key total rack solution,” said Charles Liang, president and CEO. “From the most powerful AI systems available, with up to eight NVIDIA H100 HGX GPUs to compact edge servers that must perform in challenging environmental conditions we provide the broadest portfolio of solutions for today’s most demanding workloads, including new liquid cooling solutions that reduce data center power consumption and increase performance.”
At Computex 2023, the company will be showcasing a range of servers and storage solutions and demonstrate the ntegrated rack with the newest liquid cooling technologies that enable energy efficiency and fast deployment.
Hghlights of lineup at Computex 2023 include following:
Rack scale liquid cooling – Firm’s full rack liquid cooling solution enables organizations to run the highest performing GPU servers and maintain the optimal operating conditions. The company supplies, integrates, and tests complete rack level liquid cooling solutions, including CDUs with redundant power and pumps, Cooling Distribution Manifolds (CDM), leakproof connectors, and optimized hoses. In addition, a firm’s designed highly efficient cold plate enhances heat removal from both CPUs and GPUs.
GPU-optimized systems – Servers with 8 or 4 Nvidia HGX H100 Tensor Core GPUs and dual 4th Gen Xeon Scalable processors or dual 4th Gen EPYC processors. The X13 and H13 GPU systems are open, modular, standards-based servers that provide superior performance and serviceability with a hot-swappable, toolless design. GPU options include the latest PCIe, OAM, and Nvidia SXM technology. These GPU servers are for workloads with the most demanding AI training performance, HPC, and big data analytics. The new Intel GPU Max series and a new server using the Nvidia Grace Superchip are also available.
SuperBlade – Company‘s high-performance, density-optimized, and energy-efficient X13 SuperBlade, built with 4th Gen Xeon Scalable processors, can reduce initial capital and operational expenses for many organizations. SuperBlade utilizes shared, redundant components, including cooling, networking, power, and chassis management, to deliver the compute performance of an entire server rack in a much smaller physical footprint. These systems support GPU-enabled blades and are optimized for AI, data analytics, HPC, cloud, and enterprise workloads. Compared to industry-standard servers, a cable reduction of up to 95% reduces costs and can lower power usage.
Hyper– With dual 4th Gen Xeon Scalable processors or dual 4th Gen EPYC processors – The X13 and H13 Hyper series brings next-gen performance to the firm‘s range of rackmount servers, built to take on the most demanding workloads along with the storage and IO flexibility that provides a custom fit for a range of application needs.
BigTwin (2U4N) – The X13 BigTwin with dual 4th Gen Xeon Scalable processors/node to provide superior density, performance, and serviceability with hot-swappable components in a toolless design. These systems are for cloud, storage, and media workloads.
CloudDC – With 4th Gen Xeon Scalable processors or 4th Gen EPYC processors – Ultimate flexibility on IO and storage with 2 or 6 PCIe 5.0 slots and dual AIOM slots (PCIe 5.0; OCP 3.0 compliant) for maximum data throughput. The company’s X13 and H13 CloudDC systems are designed for convenient serviceability with toolless brackets, hot-swap drive trays, and redundant power supplies that ensure rapid deployment and more efficient maintenance in data centers.
GrandTwin – Using with 4th Gen Xeon Scalable processor or 4th Gen EPYC processor – The X13 and H13 GrandTwin is purpose-built for single-processor performance. The design maximizes computing performance, memory, and efficiency to deliver maximum density. GrandTwin’s flexible modular design can be easily adapted for a range of applications, with the ability to add or remove components as required, reducing cost. In addition, the GrandTwin features front (cold aisle) hot-swappable nodes, which can be configured with either front or rear I/O for easier serviceability. As a result, the X13 and H13 GrandTwin are for workloads such as CDN, multi-access edge computing, cloud gaming, and high-availability cache clusters.
SuperEdge – Using the 4th GenXeon Scalable processor the firm‘s X13 SuperEdge is designed to handle increasing compute and IO density requirements of modern edge applications. With 3 customizable single-processor nodes, SuperEdge delivers high-class performance in a 2U, short-depth form factor. Each node is hot-swappable and offers front access I/O, making the system for remote IoT, edge, or telco deployments. Additionally, with flexible Ethernet or Fiber connectivity options to the BMC, Super Edge makes it for customers to choose remote management connections per their deployment environments.
Petascale storage – The All-Flash NVMe systems powered by a 4th Gen Xeon Scalable processor or a 4th Gen EPYC processor offer industry-leading storage density and performance with EDSFF drives, allowing capacity and performance in a single 1U chassis. The 1st in a coming lineup of X13 and H13 storage systems, these latest petascale servers support 9.5mm and 15mm E1.S or 7.5mm E3.S EDSFF (EYPC only) media designed with PCIe 5.0 interface. They are shipping from all leading flash vendors.
Liquid cooled-AI development platform
Liquid cooled-AI development platform – The deskside liquid-cooled AI development platform addresses the thermal design power needs of the 4 Nvidia A100 Tensor Core GPUs and the 2x4th Gen Xeon Scalable CPUs to enable full performance while improving the overall system’s efficiency and enabling quiet (approximately 30dB) operation in an office environment. In addition, this system is designed to accommodate high-performing CPUs and GPUs, making it for AI/DL/ML and HPC applications.
Video: Supermicro Computex 2023 Keynote – Accelerate Everything