China's "LingSheng" Exascale Supercomputer Retains Global No. 1 Ranking

2026-11-17 10:26
Favorite

en.Wedoany.com Reported - On November 16, the 68th edition of the global TOP500 supercomputer list was officially released during the SC26 International Conference for High Performance Computing, Networking, Storage, and Analysis held in Chicago, USA. The "LingSheng" exascale supercomputer deployed at the National Supercomputing Center in Shenzhen, China, retained the world No. 1 position with a measured Linpack double-precision performance of 2.198 EFlop/s, leading the second-place U.S. "El Capitan" supercomputer by approximately 21.5%, while continuing to rank first on the global HPCG list with a score of 22.00 PFlop/s.

The top five supercomputers in this edition of the TOP500 ranking are, in order: China's "LingSheng," the U.S. "El Capitan," the U.S. "Frontier," the U.S. "Aurora," and Germany's "JUPITER Booster," with corresponding measured Linpack performances of 2.198, 1.809, 1.353, 1.012, and 1.000 EFlop/s, respectively. The systems in the global top ten and their ranking order remain consistent with the June 2026 list.

The 68th edition of the list includes 27 new systems, accounting for 5.4% of all 500 systems, marking the lowest number of new entrants since the TOP500 was established in 1993. The top ten positions continue to be occupied by existing systems, with new entries in this edition mainly distributed across other ranking ranges.

On the HPCG benchmark list, "LingSheng" continues to rank first with 22.00 PFlop/s, the U.S. "El Capitan" ranks second with 17.41 PFlop/s, Japan's "Fugaku" ranks third with 16.00 PFlop/s, and the U.S. "Frontier" ranks fourth with 14.05 PFlop/s. South Korea's newly listed system "HANGANG" achieved an HPCG score of 4.29 PFlop/s, ranking eighth.

"LingSheng" was built by the Shenzhen Cloud Computing Center and adopts the domestically developed LingKun computing platform, equipped with 304-core LX2 processors, the LingQi high-speed interconnect network, and the Kylin operating system. The system is configured with 13,789,440 processor cores and has a theoretical peak computing performance of approximately 2.736 EFlop/s. The system first topped the TOP500 in June 2026, marking the return of a Chinese supercomputer to the No. 1 position after "Sunway TaihuLight" in 2017.

Technical materials previously released by the National Supercomputing Center in Shenzhen show that "LingSheng" adopts an Online Acceleration all-CPU super-intelligent fusion architecture and has supported 22 full-machine-scale applications at the tens-of-millions-of-cores level, adapting to more than 220 domain application software packages covering fields such as atmospheric and oceanic sciences, engineering simulation, materials science, and biomedicine. Large-scale first-principles computational research based on XLSDFT conducted on this system was shortlisted for the 2026 ACM Gordon Bell Prize.

This bulletin is compiled and reposted from information of global Internet and strategic partners, aiming to provide communication for readers. If there is any infringement or other issues, please inform us in time. We will make modifications or deletions accordingly. Unauthorized reproduction of this article is strictly prohibited. Email: news@wedoany.com