Skip to content
Source: theregister

🖥️China's LineShine Tops Top500 at 2.2 Exaflops, US Sits Second

The US just lost the supercomputer crown, and it won't win it back soon

TL;DR

China's LineShine tops the new Top500 list at 2.2 exaFLOPS, built despite US GPU export controls. The US is not expected to reclaim the top spot soon, but DoE's incoming Discovery system could change the math.

China's LineShine is the new number one on the Top500 supercomputer list, and it's not close. It hits 2.2 exaFLOPS on the High Performance Linpack benchmark, built by the National Supercomputing Centre in Shenzhen despite US export controls blocking its access to American GPUs. Why should you care? Because the real story here is that raw scale still wins, and the frontier of HPC is moving toward AI workloads. If you run simulation code, weather models, or large training jobs, the hardware race decides which machines your team can even rent or reach. The DoE is pushing AI-focused systems through its Genesis mission, which means more of your benchmark-driven decisions will shift toward mixed-precision and AI-friendly architectures. The numbers: LineShine draws 42.2 megawatts and runs on 45,360 LingKun LX2 CPUs, each with 304 Arm v9.2 cores at 1.5 GHz. Meanwhile, ORNL's Lux delivers 220 petaFLOPS in HPL, and its upcoming Discovery system, pairing AMD Venice CPUs with Instinct MI430X GPUs, is projected at 5.1 to 8.5 exaFLOPS FP64 Rpeak. Watch Discovery. If it lands in the projected range, the US gets back in the conversation.

China's LineShine Tops Top500 at 2.2 Exaflops, US Sits Second — theregister

Key Points

1

LineShine hits 2.2 exaFLOPS on the High Performance Linpack benchmark while drawing 42.2 megawatts of power across its full system footprint.

Keep reading — it's free

Enter your email to keep reading — plus the best of AI & tech, daily. Free, forever.

Also get
or

Already a member? Sign in

supercomputingtop500exascalehpcamd

Comments

Subscribe to join the conversation...

Be the first to comment