China's homegrown LineShine has retained its position as the world's fastest supercomputer, topping both the TOP500 and HPCG rankings released at SC26, or the International Conference for High Performance Computing, Networking, Storage and Analysis, in Chicago on Monday.
Installed at the National Supercomputing Center in the Chinese tech hub of Shenzhen, LineShine achieves 2.198 EFlops, or 2.198 quintillion floating point operations per second, on the High Performance Linpack (HPL) benchmark, according to the TOP500 organization. That puts it about 21 percent ahead of the U.S.-based El Capitan supercomputer, which ranks the second at 1.809 exaflops.
LineShine also maintains the top spot with about 22 PFlops on the High-Performance Conjugate Gradient (HPCG) benchmark .
The results mark the second consecutive first-place finish for LineShine after claiming the crown in June, and a sustained lead after a Chinese-developed supercomputer retained the top spot for the first time in nine years.
Behind the achievement is a home-developed computing architecture.
LineShine breaks with convention with a pioneering all-CPU design featuring built-in AI acceleration, backed by large-scale innovations across chips, networking, storage and system software.
It has supported 22 full-scale, ten-million-core-level applications and is compatible with over 220 domain-specific software applications.
Research achievements in the field of XLSDFT -- developed using the LineShine system -- have been shortlisted for the 2026 ACM Gordon Bell Prize, which recognizes outstanding achievement in high performance computing.
The shortlisting indicates that domestic exascale computing power has provided sound support for major, original scientific researches.
China’s LineShine remains world’s fastest supercomputer
