Researchers have developed and demonstrated what they claim to be the fastest and most powerful neuromorphic processor in the world. The processor operates faster than 10 trillion operations per second, and it is capable of processing ultralarge-scale data.
The research was led by David Moss, a professor at Swinburne University; Xinguan Xu of Swinburne and Monash Universities; and Arnan Mitchell, distinguished professor at RMIT University.
“This breakthrough was achieved with ‘optical microcombs,’ as was our world-record internet data speed reported in May 2020,” said Moss, director of Swinburne’s Optical Sciences Centre.
Microcombs are made up of hundreds of high-quality infrared lasers on a single chip and are faster, smaller, lighter, and less expensive than other optical sources.
According to Xu, the processor can serve as a universal, ultrahigh-bandwidth front end for any neuromorphic hardware, either optical or electronic, which brings large-scale data machine learning for real-time ultrahigh bandwidth data within reach.
“We’re currently getting a sneak peak of how the processors of the future will look. It’s really showing us how dramatically we can scale the power of our processors through the innovative use of microcombs,” Xu said.
Mitchell noted, “This technology is applicable to all forms of processing and communications — it will have a huge impact. Long term we hope to realize fully integrated systems on a chip, greatly reducing cost and energy consumption.”
Damien Hicks of Swinburne University and the Walter and Elizabeth Hall Institute pointed to the context and importance of the research.
“Convolutional neural networks have been central to the artificial intelligence revolution, but existing silicon technology increasingly presents a bottleneck in processing speed and energy efficiency,” Hicks said.
The research was published in Nature (www.doi.org/10.1038/s41586-020-03063-0).