The latest Top500 supercomputer ranking put China’s LineShine system in first place this week.
The machine calculated 22 per cent faster than El Capitan, the number two system at California’s Lawrence Livermore National Laboratory.
These tennis-court-sized beasts are used for frontier scientific work, including hurricane prediction and finding oil deposits. They have military uses too, the US government uses El Capitan to help maintain its nuclear weapons stockpile, among other tasks.
The US, China and Japan have fought for years over who owns the fastest machine. A Chinese system won for the first time in 2010.
In 2023, China stopped taking part in the rankings, which are published twice a year. Experts suspected a Chinese machine had already jumped ahead, but lacked the data to prove it.
This time, LineShine joined the list. The system was built by the National Supercomputing Centre in Shenzhen and unveiled in April.
LineShine relies on domestically developed central processing units rather than graphics processing units, or GPUs. That is unusual because GPUs power most of today’s leading supercomputers.
The machine uses homegrown technology for memory, networking and cooling systems too.
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 centre said.
Washington cut Chinese supercomputer developers’ access to Intel chips and other American hardware in 2015. The Biden administration later blocked China from accessing powerful GPUs, mostly developed by American chip giant Nvidia.
National Supercomputing Center in Shenzhen chief designer Lu Yutong said LineShine was designed for scientific simulations and AI workloads.
Some in the industry doubted LineShine could match dedicated AI supercomputers, which are optimised for training AI models.
