Frequency-Domain Inference Acceleration for Convolutional Neural Networks Using ReRAMs

doi:10.1109/TPDS.2023.3322907

	Frequency-Domain Inference Acceleration for Convolutional Neural Networks Using ReRAMs
	Liu, Bosheng 1; Jiang, Zhuoshen 1; Wu, Yalan 1; Wu, Jigang 1; Chen, Xiaoming 2; Liu, Peng 1; Zhou, Qingguo 3; Han, Yinhe 2
	2023-12-01
发表期刊	IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
ISSN	1045-9219
卷号	34 期号:12 页码:3133-3146
摘要	Convolutional neural networks (CNNs) (including 2D and 3D convolutions) are popular in video analysis tasks such as action recognition and activity understanding. Fast algorithms such as fast Fourier transforms (FFTs) are promising in significantly reducing computation complexity by transforming convolution into frequency domain. In frequency space, conventional spatial convolutions are replaced with simpler element-wise complex multiplications. Conventional application-specific-integrated-circuit (ASIC) based frequency-domain accelerators can achieve effective performance boost but come at the cost of significant energy consumption, owing to the hierarchical memory organization. We propose a frequency-domain resistive random access memory (ReRAM) based inference accelerator called FDA that can process element-wise complex multiplication in memory for both 2D and 3D CNNs. Each ReRAM-based frequency-domain process element (PE) with two ReRAM cells can perform an element-wise complex multiplication in two continuous execution cycles. We then provide a flexible dataflow to alleviate the redundant data movements by frequency-domain data reuse and inherent symmetrical characteristic for both 2D and 3D convolutions. Evaluation results based on representative both 2D and 3D CNN benchmarks demonstrate that FDA outperforms state-of-the-art baselines with better performance and energy efficiency.
关键词	Frequency-domain accelerator energy efficiency resistive random access memory frequency-domain convolutions
DOI	10.1109/TPDS.2023.3322907
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[62302102] ; National Natural Science Foundation of China[62122076] ; National Natural Science Foundation of China[62174038] ; State Key Laboratory of Computer Architecture (ICT, CAS)[CARCHB202119] ; Guangdong Basic and Applied Basic Research Foundation[2023A1515012844] ; Guangdong Basic and Applied Basic Research Foundation[2022A1515110599] ; Guangdong Basic and Applied Basic Research Foundation[202201010347]
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Theory & Methods ; Engineering, Electrical & Electronic
WOS记录号	WOS:001091501600004
出版者	IEEE COMPUTER SOC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/21108
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Wu, Jigang; Chen, Xiaoming
作者单位	1.Guangdong Univ Technol, Sch Comp Sci & Technol, Guangzhou 510006, Guangdong, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 3.Lanzhou Univ, Sch Informat Sci & Engn, Lanzhou 730000, Gansu, Peoples R China
推荐引用方式 GB/T 7714	Liu, Bosheng,Jiang, Zhuoshen,Wu, Yalan,et al. Frequency-Domain Inference Acceleration for Convolutional Neural Networks Using ReRAMs[J]. IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,2023,34(12):3133-3146.
APA	Liu, Bosheng.,Jiang, Zhuoshen.,Wu, Yalan.,Wu, Jigang.,Chen, Xiaoming.,...&Han, Yinhe.(2023).Frequency-Domain Inference Acceleration for Convolutional Neural Networks Using ReRAMs.IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,34(12),3133-3146.
MLA	Liu, Bosheng,et al."Frequency-Domain Inference Acceleration for Convolutional Neural Networks Using ReRAMs".IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 34.12(2023):3133-3146.