This site reported on June 26 that the world’s most powerful supercomputer is equipped with less than 50,000 super powerful GPUs. In order to enhance its competitiveness with Nvidia, AMD aims to launch a super supercomputer that supports 1.2 million GPUs. Data center AI cluster.
In an interview with TheNextPlatform, Forrest Norrod, executive vice president and general manager of AMD’s Data Center Solutions Group, said that the company is seriously considering building the largest artificial intelligenceGPU cluster. Note from this site: Artificial IntelligenceTraining clusters are usually built with thousands of GPUs, connected to server racks through high-speed interconnects, and the 1.2 million chips AMD is researching will undoubtedly increase exponentially in order of magnitude, which will inevitably bring delays, performance consumption, hardware failure and other challenges. Frontier, the world's most powerful supercomputer, "only" has 37,888 GPUs, so the artificial intelligence cluster outlined by AMD is 30 times the number of the current most powerful supercomputers.The above is the detailed content of The number is 30 times that of the most powerful supercomputer on the planet. AMD's trump card against NVIDIA: 1.2 million GPUs form an AI cluster.. For more information, please follow other related articles on the PHP Chinese website!