{"title":"边缘设备深度学习硬件加速器设计调查","authors":"Anu Samanta, Indranil Hatai, Ashis Kumar Mal","doi":"10.1007/s11277-024-11443-2","DOIUrl":null,"url":null,"abstract":"<p>In artificial intelligence, the large role is played by machine learning (ML) in a variety of applications. This article aims at providing a comprehensive survey on summarizing recent trends and advances in hardware accelerator design for machine learning based on various hardware platforms like ASIC, FPGA and GPU. In this article, we look at different architectures that allow NN executions in respect of computational units, network topologies, dataflow optimization and accelerators based on new technologies. The important features of the various strategies for enhancing acceleration performance are highlighted. The numerous current difficulties like fair comparison, as well as potential subjects and obstacles in this field has been examined. This study intends to provide readers with a fast overview of neural network compression and acceleration, a clear evaluation of different methods, and the confidence to get started in the right path.</p>","PeriodicalId":23827,"journal":{"name":"Wireless Personal Communications","volume":null,"pages":null},"PeriodicalIF":1.9000,"publicationDate":"2024-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A Survey on Hardware Accelerator Design of Deep Learning for Edge Devices\",\"authors\":\"Anu Samanta, Indranil Hatai, Ashis Kumar Mal\",\"doi\":\"10.1007/s11277-024-11443-2\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>In artificial intelligence, the large role is played by machine learning (ML) in a variety of applications. This article aims at providing a comprehensive survey on summarizing recent trends and advances in hardware accelerator design for machine learning based on various hardware platforms like ASIC, FPGA and GPU. In this article, we look at different architectures that allow NN executions in respect of computational units, network topologies, dataflow optimization and accelerators based on new technologies. The important features of the various strategies for enhancing acceleration performance are highlighted. The numerous current difficulties like fair comparison, as well as potential subjects and obstacles in this field has been examined. This study intends to provide readers with a fast overview of neural network compression and acceleration, a clear evaluation of different methods, and the confidence to get started in the right path.</p>\",\"PeriodicalId\":23827,\"journal\":{\"name\":\"Wireless Personal Communications\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":1.9000,\"publicationDate\":\"2024-07-19\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Wireless Personal Communications\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://doi.org/10.1007/s11277-024-11443-2\",\"RegionNum\":4,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"TELECOMMUNICATIONS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Wireless Personal Communications","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.1007/s11277-024-11443-2","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"TELECOMMUNICATIONS","Score":null,"Total":0}
引用次数: 0
摘要
在人工智能领域,机器学习(ML)在各种应用中发挥着重要作用。本文旨在提供一份全面的调查报告,总结基于 ASIC、FPGA 和 GPU 等各种硬件平台的机器学习硬件加速器设计的最新趋势和进展。在本文中,我们将从计算单元、网络拓扑结构、数据流优化和基于新技术的加速器等方面,探讨允许执行 NN 的不同架构。文章强调了各种提高加速性能策略的重要特点。研究还探讨了当前存在的诸多困难,如公平比较,以及该领域的潜在课题和障碍。本研究旨在为读者提供神经网络压缩和加速的快速概览、对不同方法的清晰评估,以及在正确道路上起步的信心。
A Survey on Hardware Accelerator Design of Deep Learning for Edge Devices
In artificial intelligence, the large role is played by machine learning (ML) in a variety of applications. This article aims at providing a comprehensive survey on summarizing recent trends and advances in hardware accelerator design for machine learning based on various hardware platforms like ASIC, FPGA and GPU. In this article, we look at different architectures that allow NN executions in respect of computational units, network topologies, dataflow optimization and accelerators based on new technologies. The important features of the various strategies for enhancing acceleration performance are highlighted. The numerous current difficulties like fair comparison, as well as potential subjects and obstacles in this field has been examined. This study intends to provide readers with a fast overview of neural network compression and acceleration, a clear evaluation of different methods, and the confidence to get started in the right path.
期刊介绍:
The Journal on Mobile Communication and Computing ...
Publishes tutorial, survey, and original research papers addressing mobile communications and computing;
Investigates theoretical, engineering, and experimental aspects of radio communications, voice, data, images, and multimedia;
Explores propagation, system models, speech and image coding, multiple access techniques, protocols, performance evaluation, radio local area networks, and networking and architectures, etc.;
98% of authors who answered a survey reported that they would definitely publish or probably publish in the journal again.
Wireless Personal Communications is an archival, peer reviewed, scientific and technical journal addressing mobile communications and computing. It investigates theoretical, engineering, and experimental aspects of radio communications, voice, data, images, and multimedia. A partial list of topics included in the journal is: propagation, system models, speech and image coding, multiple access techniques, protocols performance evaluation, radio local area networks, and networking and architectures.
In addition to the above mentioned areas, the journal also accepts papers that deal with interdisciplinary aspects of wireless communications along with: big data and analytics, business and economy, society, and the environment.
The journal features five principal types of papers: full technical papers, short papers, technical aspects of policy and standardization, letters offering new research thoughts and experimental ideas, and invited papers on important and emerging topics authored by renowned experts.