Floorplanning with Edge-Aware Graph Attention Network and Hindsight Experience Replay

doi:10.1145/3653453

	Floorplanning with Edge-Aware Graph Attention Network and Hindsight Experience Replay
	Bo Yang 1; Qi Xu 1; Geng H(耿浩)2 ; Song Chen 1; Bei Yu 3; Yi Kang 1
	2024-03-22
发表期刊	ACM TRANSACTIONS ON DESIGN AUTOMATION OF ELECTRONIC SYSTEMS (TODAES)
ISSN	1557-7309
EISSN	1557-7309
卷号	29 期号:5
发表状态	已发表
DOI	10.1145/3653453
摘要	In this article, we focus on chip floorplanning, which aims to determine the location and orientation of circuit macros simultaneously, so the chip area and wirelength are minimized. As the highest level of abstraction in hierarchical physical design, floorplanning bridges the gap between the system-level design and the physical synthesis, whose quality directly influences downstream placement and routing. To tackle chip floorplanning, we propose an end-to-end reinforcement learning (RL) methodology with a hindsight experience replay technique. An edge-aware graph attention network (EAGAT) is developed to effectively encode the macro and connection features of the netlist graph. Moreover, we build a hierarchical decoder architecture mainly consisting of transformer and attention pointer mechanism to output floorplan actions. Since the RL agent automatically extracts knowledge about the solution space, the previously learned policy can be quickly transferred to optimize new unseen netlists. Experimental results demonstrate that, compared with state-of-the-art floorplanners, the proposed end-to-end methodology significantly optimizes area and wirelength on public GSRC and MCNC benchmarks. © 2024 Copyright held by the owner/author(s). Publication rights licensed to ACM.
关键词	Edge aware End to end Experience replay Floor-planning Graph attention network Netlist On chips Reinforcement learnings Transformer Wire length
URL	查看原文
收录类别	SCI ; EI
语种	英语
资助项目	USTC Research Funds of the Double First-Class Initiative[YD2100002012]
WOS研究方向	Computer Science
WOS类目	Computer Science, Hardware & Architecture ; Computer Science, Software Engineering
WOS记录号	WOS:001227235300016
出版者	Association for Computing Machinery
EI入藏号	20242016105520
EI主题词	Reinforcement learning
EI分类号	723.4 Artificial Intelligence
原始文献类型	Journal article (JA)
引用统计	正在获取...
文献类型	期刊论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/357341
专题	信息科学与技术学院信息科学与技术学院_PI研究组_耿浩
作者单位	1.School of Microelectronics, University of Science and Technology of China 2.School of Information Science and Technology, ShanghaiTech University 3.Department of Computer Science and Engineering The Chinese University of Hong Kong
推荐引用方式 GB/T 7714	Bo Yang,Qi Xu,Geng H,et al. Floorplanning with Edge-Aware Graph Attention Network and Hindsight Experience Replay[J]. ACM TRANSACTIONS ON DESIGN AUTOMATION OF ELECTRONIC SYSTEMS (TODAES),2024,29(5).
APA	Bo Yang,Qi Xu,Geng H,Song Chen,Bei Yu,&Yi Kang.(2024).Floorplanning with Edge-Aware Graph Attention Network and Hindsight Experience Replay.ACM TRANSACTIONS ON DESIGN AUTOMATION OF ELECTRONIC SYSTEMS (TODAES),29(5).
MLA	Bo Yang,et al."Floorplanning with Edge-Aware Graph Attention Network and Hindsight Experience Replay".ACM TRANSACTIONS ON DESIGN AUTOMATION OF ELECTRONIC SYSTEMS (TODAES) 29.5(2024).