DNA: A General <underline>D</underline>ynamic Neural <underline>N</underline>etwork <underline>A</underline>ccelerator

doi:10.1109/TC.2025.3587617

	DNA: A General Dynamic Neural Network Accelerator
	Liu, Lian 1,2; Yu, Jinxin 1,2; Wang, Mengdi 1,2; Li, Xiaowei 1,2; Han, Yinhe 1,2; Wang, Ying 1,2
	2025-09-01
发表期刊	IEEE TRANSACTIONS ON COMPUTERS
ISSN	0018-9340
卷号	74 期号:9 页码:3210-3222
摘要	Due to the demonstrated superiority, dynamic neural networks (NNs), which adapt their network structures to different inputs, have been recognized as an optimized alternative to conventional static NNs. However, researchers have not explored the implications of dynamic NN on neural processing unit (NPU) architecture design. Consequently, we analyze the characteristics and inefficient sources of executing dynamic NNs on existing hardware. From our analysis, existing NPUs, designed for static NNs, cannot effectively handle the execution of dynamic operator and agent-dependent data loading in dynamic NNs. To this end, we present DNA, an efficient accelerator optimized to deal with the challenges of running general dynamic NNs. Firstly, to improve the execution efficiency of dynamic operators, we propose a transverter-based online scheduling strategy to rapidly generate efficient scheduling for each dynamic operator. Secondly, to mitigate hardware idleness caused by the non-deterministic and agent-dependent data access patterns in dynamic NNs, we propose a novel predictor-based prefetching strategy that achieves effective data preloading with negligible cost. We implemented our accelerator, DNA, by integrating an additional online scheduler into a typical many-core baseline accelerator. According to our evaluation of various dynamic NNs, DNA achieves 3.48x speedup and 3.03 x energy savings over the baseline accelerator.
关键词	Dynamic scheduling Artificial neural networks DNA Processor scheduling Loading Prefetching Runtime Costs Switches Optimization Dynamic NN NPU design accelerator
DOI	10.1109/TC.2025.3587617
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[62025404] ; National Natural Science Foundation of China[62222411] ; National Key R&D Program of China[2023YFB4404400]
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Hardware & Architecture ; Engineering, Electrical & Electronic
WOS记录号	WOS:001561348800003
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/2XEOYT63/41711
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Wang, Ying
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Dept Comp Sci, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Liu, Lian,Yu, Jinxin,Wang, Mengdi,et al. DNA: A General Dynamic Neural Network Accelerator[J]. IEEE TRANSACTIONS ON COMPUTERS,2025,74(9):3210-3222.
APA	Liu, Lian,Yu, Jinxin,Wang, Mengdi,Li, Xiaowei,Han, Yinhe,&Wang, Ying.(2025).DNA: A General Dynamic Neural Network Accelerator.IEEE TRANSACTIONS ON COMPUTERS,74(9),3210-3222.
MLA	Liu, Lian,et al."DNA: A General Dynamic Neural Network Accelerator".IEEE TRANSACTIONS ON COMPUTERS 74.9(2025):3210-3222.