China's LineShine supercomputer tops global rankings, surpassing US machines
China's LineShine supercomputer tops global rankings

China's LineShine Claims Top Spot in Global Supercomputer Rankings

China's LineShine supercomputer has been ranked the world's fastest in the Top500 list, surpassing the best US machines for the first time since 2017. The list, released on Tuesday, is often viewed as a measure of a nation's technological prowess.

LineShine, located at the National Supercomputing Center in Shenzhen, achieved 2.198 exaflops, meaning it can perform over 2 quintillion calculations per second. It displaced the US El Capitan supercomputer, which now ranks second. El Capitan is housed at the US government's Lawrence Livermore National Laboratory in California.

Unique Architecture of LineShine

Unlike most high-performance computers, LineShine runs entirely on conventional central processing units (CPUs) rather than graphics processing units (GPUs) commonly used for AI. According to the Top500 list, it requires approximately 42.2 megawatts of electricity to operate.

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

Supercomputers are more than 1,000 times faster than regular computers and are used for tasks such as medical breakthroughs, climate modeling, nuclear explosion simulations, human behavior prediction, and virtual weapons testing.

Global Rankings and Exascale Computers

Following El Capitan, the third and fourth spots are held by US supercomputers at national laboratories in Tennessee and Illinois. Germany's Jupiter supercomputer dropped to fifth place. These five are the only publicly verified exascale computers in the world.

Other countries with machines in the top 10 include Italy, Switzerland, and Japan. The UK has 11 machines on the list, with the University of Bristol's Isambard-AI ranked 11th, down two places. Western Australia's Setonix, ranked 86th, is the best performing of Australia's four machines.

EU Plans for AI Gigafactories

Last year, the EU unveiled a €20 billion (£17 billion) plan to build sites equipped with vast supercomputers to develop next-generation AI models. These AI "gigafactories" aim for "moonshot" innovations in healthcare, biotech, industry, robotics, and scientific discovery.

The EU strategy document states that the best-performing AI factories have up to 25,000 advanced AI processors, but a gigafactory would exceed 100,000 processors. These power-hungry facilities require huge amounts of water for cooling and should run "as much as possible" on green energy, an EU official said, with plans for water recycling. Campaigners fear these datacentres could undermine Europe's climate ambitions.

Pickt after-article banner — collaborative shopping lists app with family illustration