Jilin Zhang, Mingxuan Liang, Jinsong Wei, Shaojun Wei, Hong Chen
{"title":"A 28nm Configurable Asynchronous SNN Accelerator with Energy-Efficient Learning","authors":"Jilin Zhang, Mingxuan Liang, Jinsong Wei, Shaojun Wei, Hong Chen","doi":"10.1109/ASYNC48570.2021.00013","DOIUrl":null,"url":null,"abstract":"In this paper, we put forward an energy-efficient configurable asynchronous SNN accelerator for energy-constrained applications, which includes 256 neurons and 131K synapses with 8-bit fixed point weight. To achieve high energy efficiency and on-chip learning ability, we propose a sparse target propagation (S-TP) algorithm and design the accelerator with Click-based bundled-data asynchronous circuits. The SNN accelerator is implemented in 28nm CMOS process, and the post place and router (post-PAR) simulation results indicate that the SNN accelerator achieves on-chip learning with inference power efficiency of 3.97 pJ/SOP and 95.7% classification accuracy on NMNIST test dataset, which outperforms prior neuromorphic on-chip learning systems.","PeriodicalId":314811,"journal":{"name":"2021 27th IEEE International Symposium on Asynchronous Circuits and Systems (ASYNC)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 27th IEEE International Symposium on Asynchronous Circuits and Systems (ASYNC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ASYNC48570.2021.00013","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 7
Abstract
In this paper, we put forward an energy-efficient configurable asynchronous SNN accelerator for energy-constrained applications, which includes 256 neurons and 131K synapses with 8-bit fixed point weight. To achieve high energy efficiency and on-chip learning ability, we propose a sparse target propagation (S-TP) algorithm and design the accelerator with Click-based bundled-data asynchronous circuits. The SNN accelerator is implemented in 28nm CMOS process, and the post place and router (post-PAR) simulation results indicate that the SNN accelerator achieves on-chip learning with inference power efficiency of 3.97 pJ/SOP and 95.7% classification accuracy on NMNIST test dataset, which outperforms prior neuromorphic on-chip learning systems.