A Chinese military-developed supercomputer has once again taken first place in a global test of AI computer performance. Tianhe, created by the National University of Defense Technology, leads the list of the world’s most effective big data system, its second win since 2021.
The high computing power of supercomputers can be applied to simulations to study climate change, perform material research, explore space, and develop treatments for various diseases.
Power-efficient big data computing
The Green Graph500 list measures performance-per-watt for processing large graphs, comparing the power efficiency of fast supercomputers to handle data-intensive tasks. First published in 2010, it is the world’s leading benchmark for the performance of supercomputers in graph computation.
China’s supercomputer, the Tianhe Exa-node Prototype, reached 6,320 MTEPS/W, achieving the top position in the Green Graph500 for big data. This performance metric, measured in million horizontal edges per second per watt, evaluates the performance of the system in the communication of internal data, the South China Morning Post reports.
It surpassed the Tianhe Exa-node Prototype, which achieved 4,385 MTEPS/W and placed the first runner-up, along with Eniad, a supercomputer developed by the University of Pennsylvania, which is in third place with 2,057 MTEPS/W. Tianhe last topped the list in July 2021.
Graph analysis is a growing field in AI computing that focuses on identifying links within graph structures. It is a valuable tool for studying complex networks, patterns, and structured data. For example, it can be used to analyze social relationship maps or transaction histories in multiple accounts.
Energy-efficient big data computing is becoming increasingly important, as energy consumption is expected to be a limiting factor. As the demand for large-scale data analysis grows, graph computing is becoming an important component of big data and artificial intelligence, attracting increasing research interest worldwide.
Great success in data analysis
According to Meng Xiangfei, chief scientist at the National Supercomputer Center in Tianjin, which hosts the Tianhe supercomputer, the top ranking shows global progress in the system’s ability to handle complex data analysis, as reported by the state news agency Xinhua. The Chinese researcher added that it will play an important role in promoting the development of the next generation of intelligent technologies.
Tianhe has also contributed to the region’s major achievements in research areas such as numerical simulation technology, device computation, and weather forecasting, according to a post on the Chinese social media platform WeChat.
Furthermore, the system has also achieved major breakthroughs in industrial areas such as superintelligent integration, large-scale generative models, and supercomputing internet.
Last year, the Chinese National Supercomputing Center (NSC) in Guangzhou also introduced a domestically developed supercomputer, Tianhe Xingyi. While no details had been released, officials said it exceeded the capabilities of the famous Tianhe-2 supercomputer.