FusionServer Launched V6 Intelligent Servers

[Beijing, China, April 8, 2021] FusionServer and Intel jointly released the latest-generation FusionServer V6 intelligent servers based on the 3rd generation Intel® Xeon® Scalable processor at the press conference under the theme of "Diversified Innovations for the Ultimate Computing Power".

Kenneth Zhang from FusionServer said, "As digital transformation spreads to more industries, complex application scenarios and hybrid heavy workloads pose higher requirements on computing power. FusionServer uses innovative architecture, software, and algorithms to kick-start a generation of diversified computing."

Liang Yali, vice president of Intel Marketing Group, introduced that the 3rd-generation Intel® Xeon® Scalable processor not only ensures a balanced configuration, but also is an AI-accelerated CPU perfect for data center operations. It is designed to meet heavier and diversified computing workload requirements. No matter you use it for cloud, enterprise, AI, high-performance computing, network, or security services, this processor is optimized specifically to suit your needs.

The FusionServer V6 intelligent servers release the strongest computing power through significant upgrades and innovations in the hardware architecture, algorithms, and basic software.

· FusionServer creatively designed the rack-scale server to improve the computing density. This is a server type made for large-scale data centers, which connects L1 and L2 hardware devices to flexibly configure diversified computing, network, and storage resources, improving the computing density by four times. This is a new possibility in data center construction, where all modules can be replaced non-disruptively and O&M can be done in seconds. In doing so, the system performance is improved while the total cost of ownership (TCO) remains lowest. In terms of the system hardware, FusionServer has incorporated many proprietary designs to improve reliability, serviceability, and performance and energy efficiency, achieving 25% higher computing power than the industry average and doubling the power density.

· Besides hardware upgrade, innovations on software facilitate large-scale management. FusionServer focuses on developing state-of-the-art management software, scheduling software, and basic software. In cluster computing scenarios, FusionServer's cluster job scheduler improves the cluster scheduling capability by three times, with a maximum of 320,000 cores and up to 90% computing resource utilization. In large-scale O&M management scenarios, FusionServer servers adopt five smart technologies: deployment, discovery, upgrade, maintenance, and energy saving, to improve the O&M efficiency by 30%.

· FusionServer dwells on algorithm innovation just as well. To maximize the server performance, FusionServer provides more than 40 types of big data algorithm acceleration libraries, including machine learning and graph analysis algorithms, covering 80% service scenarios. The acceleration library ensures more than 50% higher performance than the industry average. In terms of energy saving, FusionServer developed the machine learning energy-saving algorithm (MPC-PID), which reduces energy consumption by up to 5% under the same workloads. For more stable servers, FusionServer developed the intelligent memory fault self-healing algorithm, achieving 50%+ fewer breakdowns without affecting services.

FusionServer V6 implements full-stack collaborative innovation in terms of architecture, software, and algorithms. Hardware, basic software and applications have all been upgraded for more diversified computing power. FusionServer also works with global partners to help industry customers achieve digital transformation and intelligent upgrade.