In an effort to gain ground on its rivals in the AI server market, Lenovo has launched its ThinkSystem V4 portfolio powered by Intel’s new Xeon 6 processors.
The company says these systems will increase the accessibility of AI workloads at any scale, providing options for CPU-assisted AI inference on models up to 20 billion parameters and GPU-rich modeling and training solutions for more than 175 billion parameters .
The ThinkSystem V4 portfolio includes servers optimized for rack density and transactional data processing, aiming to maximize performance in enterprise data centers, CSPs, high-performance computing and telecom companies. Lenovo says the new servers offer more than four times more cores per rack, with a 4.3x performance increase.
Neptunus liquid cooling
The new SD520 V4 server offers high rack density and efficiency, with up to 225% more cores in a 2U chassis and up to 3.18x more throughput, which should make it ideal for compute-intensive transactional workloads in industries such as online banking and e-commerce. -commerce.
The SR630 V4 is being promoted as “a new form of power for cloud scale,” with up to 42% faster media transcoding. It offers up to 2x improvement in application bandwidth thanks to PCIe 5 IO and DDR5 memory.
Lenovo also uses Neptune liquid cooling technology for the ThinkSystem V4 server range. This method has been proven to be more efficient than traditional air cooling, requiring fewer energy-consuming system fans. According to the company, liquid cooling helps meet AI performance demands while reducing energy consumption by up to 40%.
AI smarts are also being added to the new Lenovo Systems Management platform. This uses generative AI to automate and simplify server deployment and configuration.
“The Lenovo ThinkSystem V4 portfolio is designed to ensure that AI will be accessible, fair and inclusive at all stages of development and deployment,” said Kirk Skaugen (the now former) president of Lenovo Infrastructure Solutions Group.
“This next generation of Lenovo Intelligent Infrastructure delivers hybrid AI solutions that enable customers and partners to adopt AI faster and fully unleash the power of data at any workload or scale.”