The National Supercomputer Center in Shenzhen, China, announced LineShine, a new generation supercomputer system developed entirely with domestic components.

The system, which is expected to deliver 2 exaflops of performance when it reaches full capacity, was introduced at an industrial application meeting held last week.

The project aims to create an end-to-end domestic infrastructure, from processors to storage units and network technologies.

Lu Yutong, director of the Shenzhen National Supercomputer Center, emphasizes that this initiative is part of China’s strategy to reduce foreign dependence on supercomputer technologies.

High Performance with Domestic Technology

The LineShine project is being implemented through a phased development process. It is stated that in the first stage of the system, a total of 12,800 CPU cores were reached by using 100 Huawei Kunpeng servers.

In the second stage, tens of thousands of processors, large-scale interconnection infrastructure and storage systems with high data processing capacity are integrated. According to the technical documentation, the system consists of 20,480 compute nodes, and each node includes two ARMv9-based LX2 processors.

Authorities state that this architecture provides the high memory bandwidth required for scientific calculations, engineering simulations and artificial intelligence training.

Li Xiaoli from Shenzhen Science and Technology Innovation Bureau states that the project provides complete domestic control and independence at all layers.

LineShine as a Hybrid Platform

LineShine is positioned as a hybrid platform that combines artificial intelligence and scientific workloads beyond a traditional supercomputer.

System; It can be used in a wide variety of fields such as molecular simulations, fluid dynamics, material design and large-scale artificial intelligence model training.

While global supercomputer rankings are generally dominated by GPU-accelerated systems, LineShine favors a completely CPU-focused architecture.

This approach reflects China’s push to build its own high-performance computing infrastructure in the face of restrictions on access to advanced chips.

The system is designed as a long-term infrastructure investment, not just for a single generation. While the development process of the project continues, the date when the system will become fully operational has not yet been clarified.

What do you think about this ambitious supercomputer project?