TUV Rhein Joins the China ASEAN New Energy Vehicle Testing and Certification Alliance

Supermicro launches the industry's first NVIDIA HGX H100 8 and 4-GPU H100 liquid cooled servers, reducing data center power costs by 40%

The liquid cooled large-scale AI training infrastructure is a comprehensive rack integrated solution that can accelerate deployment speed, improve performance, and reduce total environmental costs

Supermicro, Inc. (NASDAQ code: SMCI) is a provider of cloud computing, AI/ML, storage, and 5G/edge integrated IT solutions, continuing to expand its data center products with liquid cooled NVIDIA HGX H100 rack solutions. Advanced liquid cooling technology is fully developed by Supermicro, which can shorten the delivery cycle of the entire equipment, improve performance, reduce operating costs, and significantly reduce data center power utilization. Compared to air-cooled data centers, data centers using Supermicro liquid cooling solutions are estimated to save 40% of power. In addition, compared to existing data centers, direct cooling costs can be reduced by 86%.

Charles Liang, President and CEO of Supermicro, stated, "Supermicro continues to lead the industry by supporting the demanding demands of global artificial intelligence workloads and modern data centers. Our innovative GPU servers, utilizing our liquid cooling technology, significantly reduce the power requirements of data centers. Based on the power required to achieve today's rapidly developing large-scale AI models, we optimize TCO and Total Environmental Cost (TCE) It is crucial for data center operators. We have proven expertise in designing and building the entire high-performance server rack. Design these GPU systems from scratch for the rack level, each using liquid cooling, with ultra-high performance and efficiency, which simplifies the deployment process. Therefore, we are able to meet customer requirements in a shorter delivery time. "

For more information about the Supermicro GPU server, please visit: htt ps::/www.supermicro. com/en/products/gpu

The AI optimized rack adopts Supermicro's latest product series, including Intel and AMD server product series, which can be quickly delivered using standard engineering templates or easily customized according to unique user requirements. Supermicro continues to provide the industry's broadest product line, providing the highest performing servers and storage systems to solve complex computationally intensive projects. Through a rack integrated solution, customers have the confidence and ability to insert and connect racks into the network, which improves efficiency compared to their own management technology.

This top-level liquid cooled GPU server includes dual Intel or AMD CPUs and 8 or 4 interconnected NVIDIA HGX H100 Tensor Core GPUs. Using liquid cooling can reduce data center power consumption by 40%, thereby reducing operational costs. In addition, both of these systems greatly surpass the previous generation NVIDIA HGX GPU device system. Compared to today's large transformer models, the performance and efficiency of the GPU GPU are improved by 30 times due to its higher interconnection speed and the use of PCIe 5.0 based network and storage.

Supermicro offers 8 NVIDIA H100 SXM5 Tensor Core GPU servers for today's largest AI model, including:

SYS-821GE-TNHR - (4th generation dual Intel Xeon scalable CPU, NVIDIA HGX H100 8 GPU, 8U) https://www.supermicro.com/en/products/system/GPU/8U/SYS-821GE-TNHR

AS-8125GS-TNHR - (4th generation dual AMD EPYC CPU, NVIDIA HGX H100 8 GPU, 8U) tps://www.supermicro.com/en/products/system/GPU/8U/AS-8125GS-TNHR

Supermicro has also designed a series of customizable GPU servers for fast AI training, extensive AI inference, or AI fusion HPC workloads, including systems equipped with four NVIDIA H100 SXM5 Tensor Core GPUs.

SYS-421GU-TNXR - (4th generation dual Intel Xeon scalable CPU, NVIDIA HGX H100 4 GPU, 4U) https://www.supermicro.com/en/products/system/GPU/4U/SYS-421GU-TNXR

SYS-521GU-TNXR - (4th generation dual Intel Xeon scalable CPU, NVIDIA HGX H100 4 GPU, 5U) https://www.supermicro.com/en/products/system/GPU/4U/SYS-521GU-TNXR

The Supermicro liquid cooled rack solution includes a coolant distribution unit (CDU), providing up to 80kW of direct chip (D2C) cooling for today's highest TDP CPU and GPU, suitable for various Supermicro servers. The power supply and liquid cooling pump are redundant and hot swappable, ensuring that the server can continue to cool in the event of a power or liquid cooling pump failure. The anti leakage connector gives customers more confidence in uninterrupted liquid cooling of all systems. For more information on the Supermicro liquid cooling system, please visit: https://www.supermicro.com/en/solutions/liquid-coolin

Rack scale design and integration have become a key service for system suppliers. As AI and HPC have become increasingly critical technologies within organizations, it is necessary to optimize and configure configurations from the server level to the entire data center to achieve maximum performance. Supermicro system and rack level experts work closely with customers to explore requirements, possess knowledge and manufacturing capabilities, and provide a large number of racks to customers worldwide.

Supermicro on ISC

To explore these technologies and meet with our experts, please visit the Supermicro booth D405 at the ISC High Performance 2023 event held in Hamburg, Germany from May 21 to 25, 2023.

About Super Micro Computer, Inc

Supermicro (NASDAQ stock code: SMCI) is a global leader in overall IT solutions for application optimization. Supermicro is established and operates in San Jose, California, dedicated to providing market innovation for enterprise, cloud, AI, and 5G telecommunications/edge IT infrastructure. We are transforming into a comprehensive IT solution provider with server, AI, storage, IoT, and switch systems, software, and services, as well as advanced high-capacity motherboards, power supplies, and chassis products. These products are designed and manufactured by companies (in the United States, Taiwan, and the Netherlands), leveraging the scale and efficiency of global operations, and optimized to improve TCO and reduce environmental impact (green computing). Server Building Block Solutions ® Our award-winning product portfolio allows customers to choose from various systems, processors, memory, GPUs, storage, networks, power, and cooling solutions (air conditioning, air cooling, or liquid cooling) built from flexible, reusable, and shaped building blocks to optimize their exact workloads and applications.