Domestic software breaks world record in molecular dynamics simulation scale

robot
Abstract generation in progress

Mars Finance News: On April 2, 2026, according to recent reports, the MatPL machine learning force field by Longxun Kuangteng leveraged a domestically produced supercomputer cluster of high-performance card supercomputers on the core nodes of the supercomputing internet. It enabled liquid water molecular dynamics simulations at a scale of 41.47 billion atoms. This achievement set a new world record for the scale of molecular dynamics simulations, making it possible— for the first time— to carry out atom-precision research on complex new materials, battery interfaces, and biological macromolecules at the 700-nanometer scale. Following the 29 billion-atom simulation result, this domestically coordinated software-and-hardware collaboration once again broke the world record, indicating that the scale limits of machine learning force field (MLFF) simulations have been substantially surpassed, while also proving the feasibility of a fully domestically developed ecosystem path for scientific simulation software in the AI era. (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