{"id":26665,"date":"2024-11-06T11:34:00","date_gmt":"2024-11-06T03:34:00","guid":{"rendered":"https:\/\/ljdevice.com.tw\/?p=26665\/"},"modified":"2024-11-06T11:34:00","modified_gmt":"2024-11-06T03:34:00","slug":"nvidia-ethernet-networking-accelerates-worlds-largest-ai-supercomputer-built-xai","status":"publish","type":"post","link":"https:\/\/ljdevice.com.tw\/en\/nvidia-ethernet-networking-accelerates-worlds-largest-ai-supercomputer-built-xai\/","title":{"rendered":"NVIDIA Ethernet Networking Accelerates World\u2019s Largest AI Supercomputer, Built by xAI"},"content":{"rendered":"

October 29, 2024<\/p>\n

\"\"<\/p>\n

NVIDIA today\u00a0announced<\/a>\u00a0that xAI\u2019s Colossus supercomputer cluster comprising 100,000 NVIDIA Hopper GPUs, achieved this massive scale by using the NVIDIA\u00a0Spectrum-X<\/a>\u00a0Ethernet networking platform, which is designed to deliver superior performance to multi-tenant, hyperscale AI factories using standards-based Ethernet, for its Remote Direct Memory Access network.<\/p>\n

Colossus, the world\u2019s largest AI supercomputer, is being used to train xAI\u2019s Grok family of large language models, with chatbots offered as a feature for X Premium subscribers. xAI is in the process of doubling the size of Colossus to a combined total of 200,000\u00a0NVIDIA Hopper<\/a>\u00a0GPUs.<\/p>\n

\u201cAI is becoming mission-critical and requires increased performance, security, scalability and cost-efficiency,\u201d\u00a0said<\/a>\u00a0Gilad Shainer<\/a>, senior vice president of networking at NVIDIA.<\/p>\n

The NVIDIA Spectrum-X Ethernet networking platform is designed to provide innovators such as xAI with faster processing, analysis and execution of AI workloads, and in turn accelerates the development, deployment and time to market of AI solutions.<\/p>\n

\u201cColossus is the most powerful training system in the world,\u201d\u00a0said<\/a>\u00a0Elon Musk<\/a>. \u201cNice work by the xAI team, NVIDIA and our many partners\/suppliers.\u201d<\/p>\n