Reconfigurable Digital Computing-In-Memory AI Chip by ECE Professors Selected as China’s 2023 Top 10 Research Advances in Semiconductors

Prof. Tu Fengbin (left) and Prof. Xie Yuan (right) have been devoted to researching and developing the groundbreaking technology of the Reconfigurable Digital Computing-In-Memory AI chip (ReDCIM). [Download Photo]

The Reconfigurable Digital Computing-In-Memory AI chip (ReDCIM) developed by Assistant Professor TU Fengbin and Fang Professor of Engineering & Chair Professor XIE Yuan, both from the Department of Electronic and Computer Engineering (ECE), was chosen as China’s 2023 Top 10 Research Advances in Semiconductors by the Journal of Semiconductors.

The top 10 entries were selected through a rigorous process conducted by a committee of 243 experts in semiconductors and integrated circuits. As the world’s first CIM-based AI chip with high-precision floating-point and integer support, ReDCIM paves the way for large-scale generative AI computing. Prof. Tu and Prof. Xie are working closely with HKUST’s AI Chip Center for Emerging Smart Systems (ACCESS), a research center under the Hong Kong government’s InnoHK initiative, for high-performance/efficiency AI chip commercialization.

Their work on ReDCIM, in collaboration with Tsinghua University, was published in IEEE Journal of Solid-State Circuits (Volume: 58, Issue: 1, January 2023), the top-1 journal in integrated circuits, with the title “ReDCIM: Reconfigurable Digital Computing-In-Memory Processor With Unified FP/INT Pipeline for Cloud AI Acceleration”. The paper was first authored by Prof. Tu, with Prof. Xie as the second last author.

Launched by the Journal of Semiconductors in 2020, the annual list of China’s Top 10 Research Advances in Semiconductors aims to recognize notable achievements in basic research on semiconductors. Over the past few years, it has gained wide recognition from experts and scholars in the semiconductor field.

Related link: