Hardware Acceleration for GCNs via Bidirectional Fusion | |
Li, Han3,4; Yan, Mingyu3,4; Yang, Xiaocheng4; Deng, Lei2; Li, Wenming4; Ye, Xiaochun4; Fan, Dongrui3,4; Xie, Yuan1 | |
刊名 | IEEE COMPUTER ARCHITECTURE LETTERS |
2021 | |
卷号 | 20期号:1页码:4 |
关键词 | Random access memory Computational modeling Analytical models Hardware Engines Computer architecture Transforms Graph convolutional neural networks hardware accelerator bidirectional execution inter-phase fusion |
ISSN号 | 1556-6056 |
DOI | 10.1109/LCA.2021.3077956 |
英文摘要 | Derived from the fusion of graph traversal and neural networks, graph convolutional neural networks (GCNs) have achieved state-of-the-art performance in graph learning. However, the hybrid execution pattern, caused by the opposite characteristics of graph traversal based phase and neural network based transformation phase, poses huge challenges to the efficient execution of traditional architectures. Although GCN accelerators have emerged to address these challenges, they fail to harvest both bidirectional execution and inter-phase fusion opportunities exposed by the alternate execution phases in GCNs. Previous works either concentrate on a single execution direction or exchange the execution order of phases without inter-phase fusion, hence failing to further improve performance and efficiency. Therefore, we propose a novel hardware unit named BiFusion, which can be easily applied to existing GCN accelerators with hybrid architecture in order to harvest both of the above opportunities. BiFusion enables dynamic direction selection and inter-phase fusion, and helps significantly reduce the amounts of data access and computation. Experiments show that integrating the BiFusion unit helps the state-of-the-art GCN accelerator achieve 2x speedup on average. |
资助项目 | National Natural Science Foundation of China[61732018] ; National Natural Science Foundation of China[61872335] ; National Natural Science Foundation of China[61802367] |
WOS研究方向 | Computer Science |
语种 | 英语 |
出版者 | IEEE COMPUTER SOC |
WOS记录号 | WOS:000658323400003 |
内容类型 | 期刊论文 |
源URL | [http://119.78.100.204/handle/2XEOYT63/17654] |
专题 | 中国科学院计算技术研究所 |
通讯作者 | Yan, Mingyu |
作者单位 | 1.Univ Calif Santa Barbara, Santa Barbara, CA 93106 USA 2.Tsinghua Univ, Beijing 100084, Peoples R China 3.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 4.Chinese Acad Sci, Inst Comp Technol, Beijing 100864, Peoples R China |
推荐引用方式 GB/T 7714 | Li, Han,Yan, Mingyu,Yang, Xiaocheng,et al. Hardware Acceleration for GCNs via Bidirectional Fusion[J]. IEEE COMPUTER ARCHITECTURE LETTERS,2021,20(1):4. |
APA | Li, Han.,Yan, Mingyu.,Yang, Xiaocheng.,Deng, Lei.,Li, Wenming.,...&Xie, Yuan.(2021).Hardware Acceleration for GCNs via Bidirectional Fusion.IEEE COMPUTER ARCHITECTURE LETTERS,20(1),4. |
MLA | Li, Han,et al."Hardware Acceleration for GCNs via Bidirectional Fusion".IEEE COMPUTER ARCHITECTURE LETTERS 20.1(2021):4. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论