Lenovo recently announced the official launch of a series of new server products based on Intel ? Xeon ?6 performance core (P-core) processors, aiming to promote the upgrade and development of data centers in the era of artificial intelligence. The new products include the Lenovo Wentian WR5220 G5, Lenovo Wentian WA5480 G5, and three servers in the Lenovo ThinkSystem V4 series, which are designed for Chinese customers to provide superior performance, flexible configuration options, and efficient cooling solutions to meet the needs of diverse AI application scenarios and accelerate the intelligent transformation of enterprises.
With the wide application of AI technology, AI development and solutions in data centers are becoming more and more diversified, and the application of small and medium-sized dedicated large models in business scenarios has gradually become the mainstream. At the same time, the data processing requirements for big data analysis and traditional applications continue to grow, which puts forward higher requirements for server performance. To meet these needs, Lenovo has partnered with Intel to launch a new server family powered by Xeon ?6 processors.
As the latest generation of Intel's products for the AI trend, the Intel ? Xeon ?6 processor adopts the advanced Intel 3 process technology, which not only achieves a significant increase in general-purpose computing power, but also makes a number of optimizations and improvements for AI applications. In particular, its performance core (P-core) design, which is optimized for compute-intensive and AI workloads, has a built-in AMX accelerator that can greatly improve the efficiency of AI-related computing tasks. This makes the Xeon ?6 processor excellent for AI inference and general-purpose AI workloads, making it a new benchmark in the field of AI computing power.
Lenovo Wentian WR5220 G5 is a 2U dual-socket rack server built for AI applications. It is equipped with Intel ? Xeon ?6 processors, each processor can have up to 86 performance cores (P-cores), which is more than 2 times more computing power than the previous generation. The WR5220 G5 also supports 6400MT/s DDR5 memory, which increases memory bandwidth by 14%, and is compatible with advanced technology standards such as CXL 2.0 and PCIe 5.0, further improving the overall performance of the system. In order to meet the complex business scenarios of medium-scale model inference, the WR5220 G5 also has flexible configuration capabilities and can be equipped with four double-wide GPUs. At the same time, the server introduces a liquid cooling solution for key components, which further reduces the PUE value of the data center and improves the energy efficiency level.
Lenovo Wentian WA5480 G5 is a 4U designed all-in-one AI server with excellent scalability and diverse computing capabilities. It is equipped with 8 high-performance AI accelerator cards, which are suitable for cutting-edge application scenarios such as deep learning, metaverse, and generative AI. The WA5480 G5 is also equipped with an Intel ? Xeon ?6 processor and adopts the new MRDIMM technology, which achieves a more than 2 times increase in memory bandwidth. The server also supports multi-brand AI accelerator cards and has a rich AI ecosystem. For high-density computing application scenarios, the WA5480 G5 also provides liquid cooling and heat dissipation design for CPU, GPU, and memory, improving heat dissipation efficiency and practicing the concept of green and low-carbon.
Lenovo's ThinkSystem V4 series servers are suitable for a wide range of workloads, from quantitative trading and web services to compute-intensive scenarios such as astrophysical research. The new ThinkSystem V4 series includes the SR630 V4, SR650 V4 and SR650a V4. Featuring high density and high space utilization, the SR630 V4 is designed for compact environments and provides superior computing power and scalability. The SR650 V4 offers maximum workload flexibility and up to 25% GPU scalability for engineering, modeling and simulation, and AI applications. Built for AI compute-intensive scenarios, the SR650a V4 supports up to four double-wide GPUs and is equipped with front-facing GPU access to ensure unmatched performance and ease of maintenance.
All three ThinkSystem V4 series servers are optimized to overcome power limitations with Lenovo's innovative Neptune liquid cooling technology. Neptune liquid cooling technology delivers up to 3.5x thermal efficiency, significantly reducing the energy required to cool IT equipment, while enhancing device processing power to support compute-intensive AI applications. The partnership between Lenovo and Intel not only helps customers scale and work smarter, but also redefines the potential of the data center by providing sustainable AI infrastructure solutions for multiple scenarios.