Compute-In-Memory (CIM) is a trending technology that fuses computation (especially for artificial intelligence) with memory integrated circuits macros. On the one hand, CIM brings emerging chip solutions for boosting AI computation efficiency by significantly reducing data transfer between processing units and memories. On the other hand, these novel array subsystem features high storage and computation density, bringing intriguing research topics of design automation for CIM integrated circuits and accelerator microarchitecture designs. Besides, CIM technology is an effective hardware acceleration solution to EDA problems. This talk will introduce the latest achievement of CIM technology, followed by the challenges and solutions of efficiently designing emerging CIM hardware.

Speaker Bio:  

Prof. Bonan Yan obtained his Ph.D degree from Department of Electrical and Computer Engineering from Duke University in 2020. He is currently an assistant professor at Institute for Artificial Intelligence, Peking University, Beijing, China. Prof. Yan’s research interests include circuits, systems and design automation for artificial intelligence chips, emerging memory integrated circuits design and in-memory computing technology. He has published over 40 research papers in top-tier journals and conferences, including ISSCC, Symposium on VLSI Technology, IEDM, DAC, ICCAD and etc. His research pioneers memristor and SRAM CIM circuits and systems design.

[1] Bonan Yan, et al., "Resistive Memory-Based In-Memory Computing: From Device and Large-Scale Integration System Perspectives," Advanced Intelligent Systems, 2019.