Full-Chip Voltage Prediction via Graph Attention Based Neural Networks

doi:10.1109/ASICON58565.2023.10396114

	Full-Chip Voltage Prediction via Graph Attention Based Neural Networks
	Li, Yuan 1; Zhou, Pingqiang2
	2023-10-24
会议录名称	2023 IEEE 15TH INTERNATIONAL CONFERENCE ON ASIC (ASICON)
ISSN	2162-7541
发表状态	已发表
DOI	10.1109/ASICON58565.2023.10396114
摘要	Power supply noise has been a rising threat to the normal functioning of a microprocessor in the form of voltage emergence. The state-of-the-art commercial chips detect such emergencies by placing a limited number of on-chip noise sensors. Current sensor-based solutions suffer from limited usable sensors, and do not work well when the power delivery network has large nonlinearity. In this paper, we propose a graph attention network-based method to predict the noises in hotspot regions. Results show that our proposed method outperforms the prior approaches (linear regression and multi-layer perception) by reducing at least 20% mean absolute error and 20% maximum absolute error on average. © 2023 IEEE.
会议录编者/会议主办者	Fudan University ; IEEE Beijing Section ; Nanjing University ; National IC Innovation Center
关键词	Power Delivery Network Voltage Prediction Noise Sensor Graph Neural Networks
会议名称	15th IEEE International Conference on ASIC, ASICON 2023
会议地点	Nanjing, China
会议日期	24-27 Oct. 2023
URL	查看原文
收录类别	EI
语种	英语
出版者	IEEE Computer Society
EI入藏号	20240715533151
EISSN	2162-755X
原始文献类型	Conference article (CA)
来源库	IEEE
文献类型	会议论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/349775
专题	信息科学与技术学院信息科学与技术学院_PI研究组_周平强组
通讯作者	Li, Yuan
作者单位	1.Duke Kunshan University, Division of Natural and Applied Sciences, Kunshan; 215316, China 2.ShanghaiTech University, School of Information Science and Technology, Shanghai; 201210, China
推荐引用方式 GB/T 7714	Li, Yuan,Zhou, Pingqiang. Full-Chip Voltage Prediction via Graph Attention Based Neural Networks[C]//Fudan University, IEEE Beijing Section, Nanjing University, National IC Innovation Center:IEEE Computer Society,2023.