|
[Invited Paper] Challenges and Trends of Memristive IMPLY-based In-memory Computing: Efficiency, Reliability, and Compatibility Perspectives
点击次数：
第一作者：Jiarui Xu
通讯作者：Chao Wang
合写作者：Boyi Dong, Mengjie Li, Yuansheng Zhao, Yuanjin Zheng and
发表刊物：IEEE International Conference on Integrated Circuits, Technologies and Applications (ICTA 2023)
收录刊物：EI
学科门类：工学
一级学科：电子科学与技术
文献类型：C
DOI码：10.1109/ICTA60488.2023.10364303
发表时间：2023-09-06
摘要：Memristive In-Memory Computing (IMC) through Material Implication logic (IMPLY) has emerged as a promising solution for intelligent edge devices. However, three major issues including computational efficiency, system reliability, and memristor array compatibility hinder the actual applications of the IMPLY-based IMC technology. In this paper, we investigate the effective solutions from system, architecture and circuit levels from the literature. Key metrics benchmark with a general Figure of Merit (FoM) including the number of memristors, operational steps, and switches are considered. Furthermore, some emerging trends of IMPLY-based IMC research are also discussed. The goal is to provide a broad perspective and a deep insight for the future research of memristive IMPLY-based IMC technology.