During his Computex 2024 keynote, Intel CEO Pat Gelsinger highlighted the progress the company has made with its Intel Xeon 6 platform and explained how it enables enterprises to refresh their aging data center systems so that they can reduce, accomplish sustainability goals, maximize physical floor and rack space, and create new digital capabilities across the enterprise.
In addition to announcing the availability of the Intel Xeon 6 E-core processors (code-named Sierra Forest), Gelsinger also said that he expected Xeon 6 P-cores (code-named Granite Rapids) to launch in Q3.
Tailored for high-density, scale-out workloads in data centers, Intel Xeon 6 E-core processors deliver significant performance improvements and power efficiency over previous generations. Notably, they enable a 3:1 rack consolidation, offering rack-level performance gains of up to 4.2 times and performance per watt improvements of up to 2.6 times compared to previous generations. This allows enterprises to refresh aging infrastructure, reduce costs, and free up space for other uses.
Gelsinger also reiterated the price-performance advantage of Intel Gaudi AI accelerators compared to the Nvidia H100 for training and inferencing large language models. Intel Gaudi 3 accelerators in an 8,192-accelerator cluster are projected to offer up to 40% faster time-to-train versus the equivalent size Nvidia H100 GPU cluster and up to 15% faster training throughput for a 64-accelerator cluster versus Nvidia H100 on the Llama2-70B model. In addition, Intel Gaudi 3 is projected to offer an average of up to 2x faster inferencing versus Nvidia H100, running popular LLMs such as Llama-70B and Mistral-7B.
As a sign of the growing momentum of the Intel Gaudi 3 accelerator, Gelsinger announced that Asus, Foxconn, Gigabyte, Inventec, Quanta and Wistron are committed to bring products to market as wells Dell, Hewlett Packard Enterprise, Lenovo and Supermicro.
Intel estimates that an AI kit comprising eight Intel Gaudi 3 accelerators with a universal baseboard (UBB) is about two-thirds the cost of comparable competitive platforms at a list price of $125,000. An equivalent Intel Gaudi 2 AI kit comes in at $65,000, one-third the cost of comparable competitive platforms.
Gadgets, gizmos, software and devices. I love all technology.
