News

NVIDIA has launched what it calls, “the world’s first Ethernet networking platform for AI” - the NVIDIA Spectrum-X Ethernet networking platform. Spectrum-X is designed to deliver the network performance needed for AI, machine learning, natural language processing, and industry applications. Comprising the NVIDIA Spectrum-4 Ethernet switch and the NVIDIA BlueField-3 SuperNIC, it leverages adaptive routing and congestion control for maximum bandwidth and noise isolation. NVIDIA claims that it offers the highest-performance Ethernet networking for AI, providing predictable outcomes for thousands of simultaneous AI jobs at every scale. The NVIDIA Spectrum SN5000 series switches are the fifth generation of Spectrum Ethernet switches and is built to accelerate hyperscale generative AI fabrics. It has port speeds reaching up to 800Gb/s to deliver accelerated Ethernet without needing to compromise between performance and features. With Spectrum-X tuned and validated across the full stack of NVIDIA hardware and software, according to NVIDIA, Spectrum-X is able to accelerate generative AI network performance by 1.6x over traditional Ethernet fabrics. To ensure that Spectrum-X isn’t just a flash in the pan, NVIDIA founder and CEO Jensen Huang announced during his COMPUTEX keynote in Taiwan that NVIDIA plans to launch new Spectrum-X products every year, delivering increased bandwidth and ports and enhanced software feature sets and programmability to keep on leading the drive for AI Ethernet networking performance. Dell Technologies, HPE, Lenovo and Supermicro are some of the partners releasing networking solutions based on the platform, with ASRock Rack, ASUS, GIGABYTE, Ingrasys, Inventec, Pegatron, QCT, Wistron and Wiwynn also their intent to release Spectrum-based products.