Sugon 60k-card AI4S Computing Cluster Deployed

robot
Abstract generation in progress

Mars Finance News, April 14: Sugon (China’s Dawning Information Industry) announced the launch of China’s largest AI4S computing cluster, with a scale of 60k cards, fully self-developed, supporting national supercomputing internet applications.
The cluster provides 2.4 EFLOPS of double-precision supercomputing power and 30 EFLOPS of half-precision AI computing power, with a total HBM capacity of 3.8 PB, total HBM bandwidth of 108 PB/s, inter-chip interconnect bandwidth of 27 PB/s, and inter-cabinet interconnect bandwidth of 3 PB/s.
With super-intelligent integration and full-precision design, it can efficiently support traditional scientific and engineering calculations (high precision) and AI large model training (low precision parallel) simultaneously. (Wide-angle observation)

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pin