SINGAPORE — China’s LineShine supercomputer claimed the No. 1 spot on the latest Top500 ranking of the world’s fastest supercomputers, outperforming the U.S.-built El Capitan by 22%, according to the ranking published this week.
The system, built by the National Supercomputing Center in Shenzhen and unveiled in April, relies entirely on domestically developed central processing units rather than graphics processing units, which power most leading supercomputers today. LineShine also uses homegrown technologies for its memory, networking, and cooling systems, the center said.
LineShine’s success “marks a historic step forward for China’s supercomputing industry in building an independent hardware and software ecosystem despite foreign technology restrictions,” the Shenzhen center said in a statement.
The U.S., China, and Japan have long competed for the top spot on the twice-yearly ranking. A Chinese machine first won in 2010. In 2023, China stopped participating in the rankings, leading experts to speculate that a Chinese machine had jumped into the lead but without the data to prove it. This time, LineShine’s operators chose to submit their benchmark results.
Washington first curtailed Chinese supercomputer developers’ access to Intel chips in 2015. The Biden administration later blocked China from accessing powerful GPUs developed by Nvidia and the tools required to produce them. The restrictions pushed Chinese developers to accelerate work on domestic alternatives.
Tennis-court-sized supercomputers are used for frontier scientific research, predicting hurricanes, and locating oil deposits. The U.S. government uses El Capitan at Lawrence Livermore National Laboratory to help maintain its nuclear-weapons stockpile, among other tasks.
While conventional supercomputers are designed for a range of problems, they generally have not been used to train leading artificial-intelligence systems. Colossus, a machine built by Elon Musk’s xAI in Tennessee with 200,000 AI chips, was more powerful by one measure than El Capitan, according to estimates published in a paper last year. Colossus did not participate in the Top500 ranking.
LineShine’s chief designer, Lu Yutong, said the system was designed to support both traditional scientific simulations and AI workloads. Some industry insiders expressed doubt that LineShine could match dedicated AI supercomputers, whose hardware is optimized for training today’s AI models, according to reports.