提高基于闪存的神经网络内存计算架构可靠性的自适应读取控制电压方案

IF 2.5 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

IEEE Transactions on Device and Materials Reliability Pub Date : 2024-07-17 DOI:10.1109/TDMR.2024.3429662

Xinrui Zhang;Jian Huang;Xianping Liu;Baiqing Zhong;Zhiyi Yu

{"title":"提高基于闪存的神经网络内存计算架构可靠性的自适应读取控制电压方案","authors":"Xinrui Zhang;Jian Huang;Xianping Liu;Baiqing Zhong;Zhiyi Yu","doi":"10.1109/TDMR.2024.3429662","DOIUrl":null,"url":null,"abstract":"The storage reliability is critical for flash memory based computing in-memory (CIM) architecture for Convolutional Neural Network (CNN). In this paper, we constructed a CIM scheme based on the Nor Flash array (NFA). We conducted simulations to investigate the impact of threshold voltage distribution and drift of Flash memory cells on the recognition accuracy for various CNN architectures based on the CIM schemes. Based on the reliability study, we proposed a novel compensation scheme to effectively mitigate the impact of threshold voltage drift and evaluated the effectiveness of the proposed scheme by recognition accuracy evaluation.","PeriodicalId":448,"journal":{"name":"IEEE Transactions on Device and Materials Reliability","volume":"24 3","pages":"422-427"},"PeriodicalIF":2.5000,"publicationDate":"2024-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"An Adaptive Read Control Voltage Scheme for Reliability Enhancement of Flash-Based In-Memory Computing Architecture for Neural Network\",\"authors\":\"Xinrui Zhang;Jian Huang;Xianping Liu;Baiqing Zhong;Zhiyi Yu\",\"doi\":\"10.1109/TDMR.2024.3429662\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The storage reliability is critical for flash memory based computing in-memory (CIM) architecture for Convolutional Neural Network (CNN). In this paper, we constructed a CIM scheme based on the Nor Flash array (NFA). We conducted simulations to investigate the impact of threshold voltage distribution and drift of Flash memory cells on the recognition accuracy for various CNN architectures based on the CIM schemes. Based on the reliability study, we proposed a novel compensation scheme to effectively mitigate the impact of threshold voltage drift and evaluated the effectiveness of the proposed scheme by recognition accuracy evaluation.\",\"PeriodicalId\":448,\"journal\":{\"name\":\"IEEE Transactions on Device and Materials Reliability\",\"volume\":\"24 3\",\"pages\":\"422-427\"},\"PeriodicalIF\":2.5000,\"publicationDate\":\"2024-07-17\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IEEE Transactions on Device and Materials Reliability\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://ieeexplore.ieee.org/document/10601194/\",\"RegionNum\":3,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"ENGINEERING, ELECTRICAL & ELECTRONIC\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Transactions on Device and Materials Reliability","FirstCategoryId":"5","ListUrlMain":"https://ieeexplore.ieee.org/document/10601194/","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}

引用次数: 0

摘要

对于基于闪存的卷积神经网络（CNN）内存计算（CIM）架构而言，存储可靠性至关重要。在本文中，我们构建了一种基于 Nor Flash 阵列（NFA）的 CIM 方案。我们进行了仿真，研究了闪存单元的阈值电压分布和漂移对基于 CIM 方案的各种 CNN 架构的识别准确率的影响。在可靠性研究的基础上，我们提出了一种新型补偿方案，以有效缓解阈值电压漂移的影响，并通过识别准确率评估来评价所提方案的有效性。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

An Adaptive Read Control Voltage Scheme for Reliability Enhancement of Flash-Based In-Memory Computing Architecture for Neural Network

The storage reliability is critical for flash memory based computing in-memory (CIM) architecture for Convolutional Neural Network (CNN). In this paper, we constructed a CIM scheme based on the Nor Flash array (NFA). We conducted simulations to investigate the impact of threshold voltage distribution and drift of Flash memory cells on the recognition accuracy for various CNN architectures based on the CIM schemes. Based on the reliability study, we proposed a novel compensation scheme to effectively mitigate the impact of threshold voltage drift and evaluated the effectiveness of the proposed scheme by recognition accuracy evaluation.

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

IEEE Transactions on Device and Materials Reliability 工程技术-工程：电子与电气

CiteScore

4.80

自引率

5.00%

发文量

审稿时长

6-12 weeks

期刊介绍： The scope of the publication includes, but is not limited to Reliability of: Devices, Materials, Processes, Interfaces, Integrated Microsystems (including MEMS & Sensors), Transistors, Technology (CMOS, BiCMOS, etc.), Integrated Circuits (IC, SSI, MSI, LSI, ULSI, ELSI, etc.), Thin Film Transistor Applications. The measurement and understanding of the reliability of such entities at each phase, from the concept stage through research and development and into manufacturing scale-up, provides the overall database on the reliability of the devices, materials, processes, package and other necessities for the successful introduction of a product to market. This reliability database is the foundation for a quality product, which meets customer expectation. A product so developed has high reliability. High quality will be achieved because product weaknesses will have been found (root cause analysis) and designed out of the final product. This process of ever increasing reliability and quality will result in a superior product. In the end, reliability and quality are not one thing; but in a sense everything, which can be or has to be done to guarantee that the product successfully performs in the field under customer conditions. Our goal is to capture these advances. An additional objective is to focus cross fertilized communication in the state of the art of reliability of electronic materials and devices and provide fundamental understanding of basic phenomena that affect reliability. In addition, the publication is a forum for interdisciplinary studies on reliability. An overall goal is to provide leading edge/state of the art information, which is critically relevant to the creation of reliable products.