Join the AI Impact Tour on June 5th in NYC to learn about auditing AI models and ensuring optimal performance and accuracy. This exclusive invite-only event will bring together top executive leaders to discuss strategies for leveraging AI effectively across organizations.
Intel has recently announced the Xeon 6 chip, designed to support the increasing demand for AI applications and models in data centers. The chip features two microarchitectures, the efficient core (E-core) and the performance core (P-core). The 6700 E-core version will launch on June 4, while the 6900 P-core version will debut in Q3 2024. Intel aims to enable customers to deliver real business outcomes through AI adoption and optimization.
In addition to the Xeon 6 chip, Intel has disclosed pricing for its Gaudi 2 and Gaudi 3 accelerator chips. The standard AI kit featuring the Gaudi 2 will cost $65,000, while the Gaudi 3 will be priced at $125,000. These accelerator chips are designed to compete with Nvidia’s H100 and offer cost-effective training and inference of large language models.
The Xeon 6 chip is positioned as a robust computing platform that excels in both performance and efficiency. It supports a wide range of use cases, from compute-intensive AI and high-performance computing to traditional enterprise applications. By modernizing their data center systems with Xeon 6 chips, companies can achieve cost savings, meet sustainability goals, optimize physical space, and generate new digital capabilities.
The Xeon 6 chip comes in two versions, the 6700 Series and the 6900 Series. The 6700 Series offers up to 144 efficient cores and 86 performance cores, while the 6900 Series offers up to 288 efficient cores and 128 performance cores. The two versions differ in terms of socket support, max TDP, memory channels, PCIe/CXL lanes, and UPI links.
Intel has chosen to stagger the release of different Xeon 6 processor SKUs based on customer demand. The E-core version is the first to launch and boasts significant benefits such as rack-level consolidation, performance gains, and energy savings. By consolidating racks from 200 to 66, companies can achieve substantial cost and space savings, as well as reduce energy consumption and carbon emissions.
The Gaudi 3 accelerator chip, scheduled for release in Q3 2024, will be supported by various system providers, including Dell, HP Enterprise, Lenovo, Supermicro, Asus, Foxconn, Gigabyte, Inventec, Quanta, and Wistron. Intel aims to help enterprises unlock the full potential of their data centers by providing efficient general compute architecture and innovative AI solutions.
The combination of Xeon and Gaudi chips is complementary, with Xeon enabling data center optimization and Gaudi accelerating specific AI workloads. The goal is to help organizations leverage large language models effectively for knowledge discovery and other use cases while achieving compatibility, performance, and value.
Intel’s comprehensive lineup of AI solutions, from semiconductor manufacturing to data center systems, positions the company as a leader in the AI market. By delivering flexible, secure, sustainable, and cost-effective solutions, Intel aims to empower customers to maximize the opportunities presented by AI.
Don’t miss out on the chance to attend The AI Impact Tour on June 5th in NYC and learn more about auditing AI models. Request your invite now and join the conversation with industry leaders.