DNN 推断中注意力计算的高性能方法和架构

Qi Cheng, Xiaofang Hu, He Xiao, Yue Zhou, Shukai Duan
{"title":"DNN 推断中注意力计算的高性能方法和架构","authors":"Qi Cheng, Xiaofang Hu, He Xiao, Yue Zhou, Shukai Duan","doi":"10.1109/TBCAS.2024.3436837","DOIUrl":null,"url":null,"abstract":"<p><p>In recent years, The combination of Attention mechanism and deep learning has a wide range of applications in the field of medical imaging. However, due to its complex computational processes, existing hardware architectures have high resource consumption or low accuracy, and deploying them efficiently to DNN accelerators is a challenge. This paper proposes an online-programmable Attention hardware architecture based on compute-in-memory (CIM) marco, which reduces the complexity of Attention in hardware and improves integration density, energy efficiency, and calculation accuracy. First, the Attention computation process is decomposed into multiple cascaded combinatorial matrix operations to reduce the complexity of its implementation on the hardware side; second, in order to reduce the influence of the non-ideal characteristics of the hardware, an online-programmable CIM architecture is designed to improve calculation accuracy by dynamically adjusting the weights; and lastly, it is verified that the proposed Attention hardware architecture can be applied for the inference of deep neural networks through Spice simulation. Based on the 100nm CMOS process, compared with the traditional Attention hardware architectures, the integrated density and energy efficiency are increased by at least 91.38 times, and latency and computing efficiency are improved by about 12.5 times.</p>","PeriodicalId":94031,"journal":{"name":"IEEE transactions on biomedical circuits and systems","volume":"PP ","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"High-Performance Method and Architecture for Attention Computation in DNN Inference.\",\"authors\":\"Qi Cheng, Xiaofang Hu, He Xiao, Yue Zhou, Shukai Duan\",\"doi\":\"10.1109/TBCAS.2024.3436837\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>In recent years, The combination of Attention mechanism and deep learning has a wide range of applications in the field of medical imaging. However, due to its complex computational processes, existing hardware architectures have high resource consumption or low accuracy, and deploying them efficiently to DNN accelerators is a challenge. This paper proposes an online-programmable Attention hardware architecture based on compute-in-memory (CIM) marco, which reduces the complexity of Attention in hardware and improves integration density, energy efficiency, and calculation accuracy. First, the Attention computation process is decomposed into multiple cascaded combinatorial matrix operations to reduce the complexity of its implementation on the hardware side; second, in order to reduce the influence of the non-ideal characteristics of the hardware, an online-programmable CIM architecture is designed to improve calculation accuracy by dynamically adjusting the weights; and lastly, it is verified that the proposed Attention hardware architecture can be applied for the inference of deep neural networks through Spice simulation. Based on the 100nm CMOS process, compared with the traditional Attention hardware architectures, the integrated density and energy efficiency are increased by at least 91.38 times, and latency and computing efficiency are improved by about 12.5 times.</p>\",\"PeriodicalId\":94031,\"journal\":{\"name\":\"IEEE transactions on biomedical circuits and systems\",\"volume\":\"PP \",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IEEE transactions on biomedical circuits and systems\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/TBCAS.2024.3436837\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE transactions on biomedical circuits and systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/TBCAS.2024.3436837","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

近年来,注意力机制与深度学习的结合在医学影像领域有着广泛的应用。然而,由于其计算过程复杂,现有的硬件架构存在资源消耗大或精度低的问题,如何将其高效地部署到 DNN 加速器上是一个难题。本文提出了一种基于内存计算(CIM)marco 的在线可编程 Attention 硬件架构,降低了 Attention 在硬件上的复杂度,提高了集成密度、能效和计算精度。首先,将Attention计算过程分解为多个级联组合矩阵运算,以降低其在硬件端的实现复杂度;其次,为了降低硬件非理想特性的影响,设计了一种在线可编程CIM架构,通过动态调整权重来提高计算精度;最后,通过Spice仿真验证了所提出的Attention硬件架构可应用于深度神经网络推理。基于 100nm CMOS 工艺,与传统的 Attention 硬件架构相比,集成密度和能效至少提高了 91.38 倍,延迟和计算效率提高了约 12.5 倍。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
High-Performance Method and Architecture for Attention Computation in DNN Inference.

In recent years, The combination of Attention mechanism and deep learning has a wide range of applications in the field of medical imaging. However, due to its complex computational processes, existing hardware architectures have high resource consumption or low accuracy, and deploying them efficiently to DNN accelerators is a challenge. This paper proposes an online-programmable Attention hardware architecture based on compute-in-memory (CIM) marco, which reduces the complexity of Attention in hardware and improves integration density, energy efficiency, and calculation accuracy. First, the Attention computation process is decomposed into multiple cascaded combinatorial matrix operations to reduce the complexity of its implementation on the hardware side; second, in order to reduce the influence of the non-ideal characteristics of the hardware, an online-programmable CIM architecture is designed to improve calculation accuracy by dynamically adjusting the weights; and lastly, it is verified that the proposed Attention hardware architecture can be applied for the inference of deep neural networks through Spice simulation. Based on the 100nm CMOS process, compared with the traditional Attention hardware architectures, the integrated density and energy efficiency are increased by at least 91.38 times, and latency and computing efficiency are improved by about 12.5 times.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
An Energy-Efficient and Artifact-Resilient ASIC for Simultaneous Neural Recording and Optogenetic Stimulation. Integrated Real-Time CMOS Luminescence Sensing and Impedance Spectroscopy in Droplet Microfluidics. Dynamic sub-array selection-based energy-efficient localization and tracking method to power implanted medical devices in scattering heterogenous media employing ultrasound. A Reconfigurable Bidirectional Wireless Power and Full-Duplex Data Transceiver IC for Wearable Biomedical Applications. An Ultrasonic Transceiver for Non-Invasive Intracranial Pressure Sensing.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1