-+ 0.00%
-+ 0.00%
-+ 0.00%

On November 17, local time, during the SC26 International Supercomputing Conference held in Chicago, USA, a new edition of the global supercomputing TOP500 list was officially released: the “Ling Sheng” e-level supercomputer of the National Supercomputing Center ranked first in the world with Linpack's double-precision measured performance of 2.198 eFlops, leading the second-place US “El Capitan” by about 21%, and continuing to top the HPCG list with 22 pflops. After reaching the top for the first time in June of this year, China once again took first place in the world in supercomputing, and is also a steady leader after climbing the top of supercomputing after a lapse of nine years. The system pioneered an ultra-intelligent integrated architecture for online acceleration of all CPUs, enabling autonomous control of the full stack from chips, networks, and storage to system software. It has supported 22 multi-core full-scale applications, and is suitable for more than 220 applications in more than 220 fields. XLSDFT research and application results based on “Ling Sheng” were shortlisted for the 2026 ACM Gordon Bell Award, demonstrating the ability of domestic E-class computing power to support major original research.

智通财经·11/17/2026 02:01:08
语音播报
On November 17, local time, during the SC26 International Supercomputing Conference held in Chicago, USA, a new edition of the global supercomputing TOP500 list was officially released: the “Ling Sheng” e-level supercomputer of the National Supercomputing Center ranked first in the world with Linpack's double-precision measured performance of 2.198 eFlops, leading the second-place US “El Capitan” by about 21%, and continuing to top the HPCG list with 22 pflops. After reaching the top for the first time in June of this year, China once again took first place in the world in supercomputing, and is also a steady leader after climbing the top of supercomputing after a lapse of nine years. The system pioneered an ultra-intelligent integrated architecture for online acceleration of all CPUs, enabling autonomous control of the full stack from chips, networks, and storage to system software. It has supported 22 multi-core full-scale applications, and is suitable for more than 220 applications in more than 220 fields. XLSDFT research and application results based on “Ling Sheng” were shortlisted for the 2026 ACM Gordon Bell Award, demonstrating the ability of domestic E-class computing power to support major original research.