Chinese supercomputer leapfrogs best US machines to be ranked world’s fastest | Computing
A supercomputer in China now surpasses its U.S. counterpart as the world’s most powerful. For the first time since 2017, a Chinese computer tops a list sometimes seen as a measure of a country’s technological might.
The LineShine computer in Shenzhen replaced El Capitan, the top-ranked US computer, in the Top500 rankings released on Tuesday. This was LineShine’s debut on the chart.
China’s LineShine differs from other high-performance computers in that it runs entirely on traditional computer chips (CPUs) rather than the graphics processors (GPUs) commonly used for artificial intelligence. According to the listing, it requires approximately 42.2 megawatts of electricity to operate.
Supercomputers, which are 1,000 times faster than an ordinary computer, can be used to research medical breakthroughs, model climate systems, simulate nuclear explosions, predict human behavior and conduct virtual weapons tests.
Scientists participating in the Top500 project said that LineShine at the China National Supercomputing Center achieved 2,198 exaflops, which can perform more than 2 quintillion calculations per second.
El Capitan, located at the US government’s Lawrence Livermore National Laboratory in California, is second only to two other US supercomputers located at national laboratories in Tennessee and Illinois.
In fifth place is the Jupiter supercomputer in Germany. The five are the only publicly verified exascale computers in the world.
Other countries where machines rank in the top 10 include Italy, Switzerland and Japan.
The UK has 11 machines on the 500-machine list. The University of Bristol’s Isambard-AI tops this group, falling two places since the last rankings to 11th. Equipped with 5,400 Nvidia “superchips,” Isambard-AI sits inside a black metal cage covered with barbed wire.
Western Australia’s Setonix (ranked 86th) is the best performing machine of the four available in Australia.
As Europe tries to catch up with leaders in the US and China, last year the EU unveiled a €20bn (£17bn) plan to build sites equipped with massive supercomputers to develop next-generation artificial intelligence models.
AI “gigafactories” will target “moonshot” innovations in fields such as healthcare, biotechnology, industry, robotics and scientific discovery.
The EU strategy document said the best-performing AI factories have supercomputers equipped with up to 25,000 advanced AI processors, but a gigafactory would exceed 100,000 AI processors.
An EU official said these energy-hungry plants, which can require large amounts of water for cooling, should run on a green energy source “as much as possible”, with plans to “recycle” the water if used.
Campaigners fear power-hungry data centers could undermine Europe’s climate targets.
Additional reporting by Jennifer Rankin, Robert Booth and the Associated Press
