According to the Institute for Artificial Intelligence at Peking University, a research team led by Researcher Sun Zhong and his collaborators has recently published a paper in the international journal Nature Electronics, reporting a major breakthrough in novel computing architectures.
The team successfully developed a high-precision and scalable analog matrix computing chip based on resistive random-access memory (RRAM). For the first time, the chip achieves analog computation accuracy rival to that of digital systems, improving the precision of traditional analog computing by an astonishing five orders of magnitude.
Performance evaluations show that when solving large-scale MIMO signal detection and other key scientific problems, the chip’s computational throughput and energy efficiency are hundreds to thousands of times higher than those of today’s top-tier digital processors (GPU).