Chinese Scientists Develop Brain-Mimicking Chip Claimed to Outperform Nvidia GPU
Translated from English, summarized and contextualized by DistantNews.
At a glance
- Chinese scientists developed a computer chip that models brain structures in real time, reportedly up to 478 times faster than Nvidia's A100 GPU.
- The chip integrates an artificial neural network and can reconstruct complex brain surfaces in under half a second.
- Developers believe the technology could revolutionize diagnostics for conditions like Alzheimer's, enhance brain-machine interfaces, and aid surgical navigation.
Chinese researchers have unveiled a novel computer chip capable of real-time brain structure modeling, a development they claim significantly outperforms current leading technology. Developed by scientists from Peking University and the Chinese Academy of Sciences, the chip boasts speeds up to 478 times faster than Nvidia's state-of-the-art A100 graphics processing unit (GPU).
The breakthrough, detailed in a peer-reviewed study published in Science, centers on a 40-nanometer memory chip integrated with an artificial neural network. This design overcomes previous computational limitations, allowing it to reconstruct intricate brain surfaces in less than half a second. This speed is crucial for applications requiring immediate data processing.
This breakthrough opens up new possibilities for brain-computer interfaces and the diagnosis and treatment of brain diseases. In the future, personalised and dynamic digital brain twins will become possible.
Lead author Yang Yuchao, a professor at Peking University, stated that the chip's ability to accurately render brain folds opens new avenues for medical applications. "This breakthrough opens up new possibilities for brain-computer interfaces and the diagnosis and treatment of brain diseases," Yang told the state-run Guangming Daily. He envisions a future where personalized and dynamic digital brain twins become a reality.
The technology also promises to provide a real-time hardware foundation for intraoperative neuronavigation, aiding surgeons during procedures. Furthermore, it could facilitate early screening for Alzheimer's disease and enable personalized interventions, offering new hope for patients and medical professionals alike.
It also provides a hardware foundation that can operate in real time for intraoperative neuronavigation, early screening for Alzheimerโs disease and personalised interventions.
Originally published by South China Morning Post in English. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.