A near-threshold memristive computing-in-memory engine for edge intelligence

doi:10.1038/s41467-025-61025-4

	A near-threshold memristive computing-in-memory engine for edge intelligence
	Wang, Linfang 1,2; Li, Weizeng 1,3; Zhou, Zhidao 1,3; An, Junjie 1; Ye, Wang 1,3; Li, Zhi 1,3; Gao, Hanghang 1,3; Hu, Hongyang 1,3; Liu, Jing 1,3; Chen, Xiaoming 4; Li, Ling 5; Liu, Qi 1,6; Seok, Mingoo 2; Dou, Chunmeng 1,3; Liu, Ming 1,6
	2025-07-01
发表期刊	NATURE COMMUNICATIONS
卷号	16 期号:1 页码:10
摘要	Memristive computing-in-memory and near-threshold computing are two unconventional computing paradigms that can potentially enhance the energy efficiency and real-time performance of edge devices. However, their scalability faces challenges, primarily due to process variation. Here, we report a 1-Mb, 16-macro near-threshold memristive computing-in-memory engine. The two-transistor-one-resistor cells provide strong cell current modulation capability with more than 120-times amplified resistance ratio. To mitigate variation issues, we compensate for transistor mismatches by leveraging the intrinsic variations in memristors. Additionally, we propose a charge stacking technique between multiple analog-to-digital converters to perform analog weight-and-combine operations with small energy and area overhead. Moreover, we introduce an inter-macro hybrid control scheme to reduce the task-level inference power. The fabricated chip can perform highly parallel analog computing over 256 input channels with a 2.4% relative standard deviation. It achieves a throughput up to 10.49 tera-operations per second and an energy efficiency up to 88.51 tera-operations per second per watt.
DOI	10.1038/s41467-025-61025-4
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China (NSFC)[92364202] ; National Natural Science Foundation of China (NSFC)[62488101] ; National Natural Science Foundation of China (NSFC)[U2441247] ; National Natural Science Foundation of China (NSFC)[62425407]
WOS研究方向	Science & Technology - Other Topics
WOS类目	Multidisciplinary Sciences
WOS记录号	WOS:001523451900045
出版者	NATURE PORTFOLIO
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/42013
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Dou, Chunmeng
作者单位	1.Chinese Acad Sci, Inst Microelect, State Key Lab Fabricat Technol Integrated Circuits, Beijing, Peoples R China 2.Columbia Univ, Dept Elect Engn, New York, NY USA 3.Univ Chinese Acad Sci, Beijing, Peoples R China 4.Chinese Acad Sci, Inst Comp Technol, Beijing, Peoples R China 5.Chinese Acad Sci, Inst Software, Beijing, Peoples R China 6.Fudan Univ, Frontier Inst Chip & Syst, Shanghai, Peoples R China
推荐引用方式 GB/T 7714	Wang, Linfang,Li, Weizeng,Zhou, Zhidao,et al. A near-threshold memristive computing-in-memory engine for edge intelligence[J]. NATURE COMMUNICATIONS,2025,16(1):10.
APA	Wang, Linfang.,Li, Weizeng.,Zhou, Zhidao.,An, Junjie.,Ye, Wang.,...&Liu, Ming.(2025).A near-threshold memristive computing-in-memory engine for edge intelligence.NATURE COMMUNICATIONS,16(1),10.
MLA	Wang, Linfang,et al."A near-threshold memristive computing-in-memory engine for edge intelligence".NATURE COMMUNICATIONS 16.1(2025):10.