一种用于视觉芯片的8t内存SRAM单元像素并行阵列处理器

Leyi Chen, Junxian He, Jianyi Yu, Haibing Wang, Jing Lu, Liyuan Liu, N. Wu, Cong Shi, Tian Min
{"title":"一种用于视觉芯片的8t内存SRAM单元像素并行阵列处理器","authors":"Leyi Chen, Junxian He, Jianyi Yu, Haibing Wang, Jing Lu, Liyuan Liu, N. Wu, Cong Shi, Tian Min","doi":"10.1109/APCCAS55924.2022.10090359","DOIUrl":null,"url":null,"abstract":"Vision chip is a high-speed image processing device, featuring a massively-parallel pixel-level processing element (PE) array to boost pixel processing speed. However, the collocated processing unit and fine-grained data memory unit inside each PE impose a huge requirement on memory access bandwidth as well as big area and energy consumption. To overcome this bottleneck, this paper proposes a full custom $\\mathbf{8T}$ SRAM-based Processing-in-memory (PIM) architecture to realize pixel-parallel array processor for high-speed energy-efficient vision chips. The proposed PIM architecture is constructed by emending multiplexer-based computing circuits into a dual port 8T SRAM array, so as to form a PIM PE array. Each PIM PE holds a 66-bit 8T SRAM cell block embedding in-memory logic functions, of which 64-bit 8T SRAM cells serving as the PE memory, 2-bit 8T SRAM cells acting as a buffer register in the PE. A full custom physical layout of a 16 $\\times \\boldsymbol{16}$ prototyping PIM PE array is designed and simulated using a 65 nm CMOS technology. The simulation results demonstrate that our proposed PIM PE architecture can achieve 200 MHz operation at 1.2 V, and reach a high energy efficiency of 3.97 TOPS/W while keeping a compact area of 0.129 $\\mathbf{mm}^{2}$.","PeriodicalId":243739,"journal":{"name":"2022 IEEE Asia Pacific Conference on Circuits and Systems (APCCAS)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"An 8-T Processing-in-Memory SRAM Cell-Based Pixel-Parallel Array Processor for Vision Chips\",\"authors\":\"Leyi Chen, Junxian He, Jianyi Yu, Haibing Wang, Jing Lu, Liyuan Liu, N. Wu, Cong Shi, Tian Min\",\"doi\":\"10.1109/APCCAS55924.2022.10090359\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Vision chip is a high-speed image processing device, featuring a massively-parallel pixel-level processing element (PE) array to boost pixel processing speed. However, the collocated processing unit and fine-grained data memory unit inside each PE impose a huge requirement on memory access bandwidth as well as big area and energy consumption. To overcome this bottleneck, this paper proposes a full custom $\\\\mathbf{8T}$ SRAM-based Processing-in-memory (PIM) architecture to realize pixel-parallel array processor for high-speed energy-efficient vision chips. The proposed PIM architecture is constructed by emending multiplexer-based computing circuits into a dual port 8T SRAM array, so as to form a PIM PE array. Each PIM PE holds a 66-bit 8T SRAM cell block embedding in-memory logic functions, of which 64-bit 8T SRAM cells serving as the PE memory, 2-bit 8T SRAM cells acting as a buffer register in the PE. A full custom physical layout of a 16 $\\\\times \\\\boldsymbol{16}$ prototyping PIM PE array is designed and simulated using a 65 nm CMOS technology. The simulation results demonstrate that our proposed PIM PE architecture can achieve 200 MHz operation at 1.2 V, and reach a high energy efficiency of 3.97 TOPS/W while keeping a compact area of 0.129 $\\\\mathbf{mm}^{2}$.\",\"PeriodicalId\":243739,\"journal\":{\"name\":\"2022 IEEE Asia Pacific Conference on Circuits and Systems (APCCAS)\",\"volume\":\"45 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-11-11\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2022 IEEE Asia Pacific Conference on Circuits and Systems (APCCAS)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/APCCAS55924.2022.10090359\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 IEEE Asia Pacific Conference on Circuits and Systems (APCCAS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/APCCAS55924.2022.10090359","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1

摘要

视觉芯片是一种高速图像处理器件,采用大规模并行像素级处理元件(PE)阵列来提高像素处理速度。然而,每个PE内部的并发处理单元和细粒度数据存储单元对内存访问带宽提出了巨大的要求,并且占用了较大的面积和能耗。为了克服这一瓶颈,本文提出了一种基于全定制$\mathbf{8T}$ sram的内存处理(PIM)架构,以实现用于高速节能视觉芯片的像素并行阵列处理器。提出的PIM架构是将基于多路复用器的计算电路修正为双端口8T SRAM阵列,从而形成PIM PE阵列。每个PIM PE持有一个嵌入内存逻辑功能的66位8T SRAM单元块,其中64位8T SRAM单元作为PE内存,2位8T SRAM单元作为PE中的缓冲寄存器。采用65纳米CMOS技术,设计并模拟了一个完整的自定义物理布局的16 $\times \boldsymbol{16}$原型PIM PE阵列。仿真结果表明,我们提出的PIM PE架构可以在1.2 V下实现200 MHz的工作,并达到3.97 TOPS/W的高能效,同时保持0.129 $\mathbf{mm}^{2}$的紧凑面积。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
An 8-T Processing-in-Memory SRAM Cell-Based Pixel-Parallel Array Processor for Vision Chips
Vision chip is a high-speed image processing device, featuring a massively-parallel pixel-level processing element (PE) array to boost pixel processing speed. However, the collocated processing unit and fine-grained data memory unit inside each PE impose a huge requirement on memory access bandwidth as well as big area and energy consumption. To overcome this bottleneck, this paper proposes a full custom $\mathbf{8T}$ SRAM-based Processing-in-memory (PIM) architecture to realize pixel-parallel array processor for high-speed energy-efficient vision chips. The proposed PIM architecture is constructed by emending multiplexer-based computing circuits into a dual port 8T SRAM array, so as to form a PIM PE array. Each PIM PE holds a 66-bit 8T SRAM cell block embedding in-memory logic functions, of which 64-bit 8T SRAM cells serving as the PE memory, 2-bit 8T SRAM cells acting as a buffer register in the PE. A full custom physical layout of a 16 $\times \boldsymbol{16}$ prototyping PIM PE array is designed and simulated using a 65 nm CMOS technology. The simulation results demonstrate that our proposed PIM PE architecture can achieve 200 MHz operation at 1.2 V, and reach a high energy efficiency of 3.97 TOPS/W while keeping a compact area of 0.129 $\mathbf{mm}^{2}$.
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
×
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
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
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 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信