Inspur Information G7 server fully supports the fifth generation Intel ® xeon ® Scalable processor

Through innovative design in system architecture, hardware, firmware, and cooling, as well as the introduction of AI acceleration features, the general computing performance of G7 servers has been improved by 23%, with over 36% improvement in per watt characteristics in customer workloads, especially in AI inference scenarios, with a performance improvement of up to 42%. This achieves the best performance release per unit power consumption and better meets customer requirements for various computationally intensive workloads, providing AIGC, cloud computing, virtualization Database analysis, quantitative trading and other application scenarios provide superior performance.

More efficient, powerful, and suitable for AI applications

The common demand for computing power in various fields of society is showing explosive growth, and artificial intelligence applications represented by AIGC are rapidly rising, putting higher demands on computing power. Equipped with the fifth generation Intel ® xeon ® The processor's Wave Information G7 server has increased the maximum number of cores to 64, with an average performance improvement of up to 23% compared to the previous generation product. It supports full line PCIe 5.0, 4 sets of CPU UPI interconnection configurations, and the speed has been increased to 20GT/s. It also supports DDR5 5600, which increases the memory bandwidth by 17% compared to 4800, allowing the platform's computing power to be maximized.

In order to run AI applications more efficiently, based on the fifth generation Xeon ® The processor and G7 server use AI acceleration technology to achieve an improvement of over 42% in AI inference ability. The cross CPU UPI interconnection speed of the G7 server is increased by 25%, significantly improving cross path data processing efficiency, reducing business access latency, and supporting fine-grained power consumption adjustment. Its new three-level caching function has increased the cache capacity of the G7 server by three times, allowing data to be saved in on-chip cache with higher bandwidth and lower latency, Reduce data overflow to memory and hard drives for processing, effectively optimize business latency, improve data access efficiency by 20% -30%, and meet the demand for high-speed data processing in AI reasoning applications such as voice and image recognition.

The Industry's Most Rich Adaptation to Intel's Fifth Generation Xeon ® Product portfolio of processors

The diversification of enterprise application scenarios such as financial transactions, e-commerce platforms, online games, XR interaction, and 8K videos has driven the upgrading of computing power. The widespread application of artificial intelligence not only brings new opportunities to the computing power market, but also puts forward higher and more diversified requirements. Faced with the diverse needs of customer business scenarios, G7 servers, based on the specific requirements of deployment and application environments, have launched various types of products such as computing density optimization, storage density optimization, IO expansion optimization, and full cabinet deployment optimization to cater to the differences in application load and scenario diversification, providing customers with the best computing application platform.

This wave information G7 computing power platform upgrade supports 12 server products, including NF5180G7, NF5280G7, NF5266G7, i24G7, NF5468G7, NF5688G7, etc., providing customers with customized product solutions.

I24G7: A new generation high-density multi node server that supports 4 dual nodes in 2U space, with a total of 8 5th generation Intel nodes ® Xeon ® Scalable processor. Wind liquid combination, efficient liquid cooling, integrated cold plate design, covering CPU, memory, and VR, supporting 45 ℃ medium and high temperature water inlet cooling, meeting the low PUE requirements of the system.

NF5180G7: High density computing type 1U2S universal server, 2U supports 2 5th generation Intel chips ® Xeon ® Scalable processor, 32 PCIe 5.0 E1. S direct connected storage matrices, industry's highest performance parallel storage, imported air-cooled EVAC, cold plate liquid cooling, and immersive liquid cooling multi-dimensional cooling solutions, to meet the low PUE demands of more high-density data centers.

NF5266G7: High density storage server, the only one in the industry that can deploy 28 3.5-inch large capacity hard drives in 2U space, with a storage capacity of over 600T, supporting 2 fifth generation Intel ® Xeon ® Scalable processor, standard PCIe 5.0 with a speed of 32GT/s, supporting 400Gb networks, achieving low latency application experience, better memory to computing ratio, and flexibly meeting the needs of various big data application scenarios.

NF5280G7: Versatile flagship 2U2S universal server, 2U supports 2 5th generation Intel chips ® Xeon ® Scalable processor, up to 32 DDR5 memory, 20 PCI-E extensions, 24 E3-S SSDs, fully supporting platinum+/titanium power supplies, with an optimal power conversion efficiency of up to 98%, meeting the needs of different scenarios from general computing to heterogeneous computing, storage to IO, achieving "versatile design, one machine for all".

NF5468G7: An AI server with both training and inference capabilities, equipped with two fifth generation Intel chips in a 4U space ® Xeon ® Scalable processors, 8 extreme performance GPU accelerators, and flexible high-speed interconnected networks provide up to 128 processor cores and 8TB of system memory, meeting complex application scenarios such as large models, AIGC, deep learning, metaverse, and AI+Science.

NF5688G7: Leading edge AI server with 6U space supporting 2 5th generation Intel chips ® Xeon ® Scalable processor with 8 leading high-speed interconnect acceleration chips, supporting up to 4.0 Tbps network bandwidth and 128TB NVMe high-speed storage, meeting the strong performance and ultimate scalability requirements of ultra large scale data centers, suitable for large model pre training, AIGC and other application scenarios.

With the acceleration of intelligent transformation in various industries, Inspur Information, as a leading global provider of computing infrastructure, will always adhere to system design as the core, and adhere to the design concept of openness, diversity, extreme, and green to build computing infrastructure. Currently, G7 servers have been widely used in industries such as the Internet, energy, transportation, and healthcare. Recently, according to the "Global Server Market Tracking Report for the Third Quarter of 2023" released by international data company IDC, Inspur Information Server ranked second globally.