A 22nm 3.5TOPS/W Flexible Micro-Robotic Vision SoC with 2MB eMRAM for Fully-on-Chip Intelligence

We present a highly flexible micro-robotic vision SoC featuring a hybrid Processing Element (PE) for efficient processing of both Convolutional Neural Network (CNN) and non-CNN vision tasks with 2MB embedded MRAM for retentive fully-on-chip weight storage. Fabricated in 22nm, the design achieves 0.2...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhang, Qirui, An, Hyochan, Fan, Zichen, Wang, Zhehong, Li, Ziyun, Wang, Guanru, Kim, Hun-Seok, Blaauw, David, Sylvester, Dennis
Format: Conference Proceeding
Language:English
Subjects:
Convolutional neural networks
Corner detection
Task analysis
Very large scale integration
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page 73
container_issue
container_start_page 72
container_title
container_volume
creator Zhang, Qirui
An, Hyochan
Fan, Zichen
Wang, Zhehong
Li, Ziyun
Wang, Guanru
Kim, Hun-Seok
Blaauw, David
Sylvester, Dennis
description We present a highly flexible micro-robotic vision SoC featuring a hybrid Processing Element (PE) for efficient processing of both Convolutional Neural Network (CNN) and non-CNN vision tasks with 2MB embedded MRAM for retentive fully-on-chip weight storage. Fabricated in 22nm, the design achieves 0.22nJ/pix for Harris corner detection (a non-CNN vision task) and 3.5TOPS/W (INT16) for CNN, a 60% efficiency improvement over state-of-the-art NVM-based NN ASICs.
doi_str_mv 10.1109/VLSITechnologyandCir46769.2022.9830340
format conference_proceeding
fullrecord <record><control><sourceid>ieee_CHZPO</sourceid><recordid>TN_cdi_ieee_primary_9830340</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>9830340</ieee_id><sourcerecordid>9830340</sourcerecordid><originalsourceid>FETCH-LOGICAL-i118t-b964388cc613ea85097a877006f6d3ac1be45b378f494e7e70b35b7786fcd0d23</originalsourceid><addsrcrecordid>eNot0E1PwjAcgPFqYiIin8BLv8BmX9a3Iy6iJCwYQDzi2v0HNaUl24zy7T3I6bn9Dg9COSU5pcQ8bhfr-QbcIaaQ9uc6NqXvCqmkyRlhLDeaE16QKzQxSlMpRWGUYuIajRgVOjNSs1t01_dfhDAiuB6hzylmLB4xz8Vm-bZ-_MCzAL_eBsCVd13KVsmmwTu89b1PEa9TiX_8cMCsesJQraYVblOHZ98hnLMUs_LgT3geBwjB7yE6uEc3bR16mFw6Ru-z5035mi2WL_Nyusg8pXrIrJEF19o5STnUWhCjaq0UIbKVDa8dtVAIy5VuC1OAAkUsF1YpLVvXkIbxMXr4dz0A7E6dP9bdeXcZwv8AnM1Y7Q</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>A 22nm 3.5TOPS/W Flexible Micro-Robotic Vision SoC with 2MB eMRAM for Fully-on-Chip Intelligence</title><source>IEEE Xplore All Conference Series</source><creator>Zhang, Qirui ; An, Hyochan ; Fan, Zichen ; Wang, Zhehong ; Li, Ziyun ; Wang, Guanru ; Kim, Hun-Seok ; Blaauw, David ; Sylvester, Dennis</creator><creatorcontrib>Zhang, Qirui ; An, Hyochan ; Fan, Zichen ; Wang, Zhehong ; Li, Ziyun ; Wang, Guanru ; Kim, Hun-Seok ; Blaauw, David ; Sylvester, Dennis</creatorcontrib><description>We present a highly flexible micro-robotic vision SoC featuring a hybrid Processing Element (PE) for efficient processing of both Convolutional Neural Network (CNN) and non-CNN vision tasks with 2MB embedded MRAM for retentive fully-on-chip weight storage. Fabricated in 22nm, the design achieves 0.22nJ/pix for Harris corner detection (a non-CNN vision task) and 3.5TOPS/W (INT16) for CNN, a 60% efficiency improvement over state-of-the-art NVM-based NN ASICs.</description><identifier>EISSN: 2158-9682</identifier><identifier>EISBN: 9781665497725</identifier><identifier>EISBN: 1665497726</identifier><identifier>DOI: 10.1109/VLSITechnologyandCir46769.2022.9830340</identifier><language>eng</language><publisher>IEEE</publisher><subject>Convolutional neural networks ; Corner detection ; Task analysis ; Very large scale integration</subject><ispartof>2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits), 2022, p.72-73</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/9830340$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,777,781,786,787,27906,54536,54913</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/9830340$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Zhang, Qirui</creatorcontrib><creatorcontrib>An, Hyochan</creatorcontrib><creatorcontrib>Fan, Zichen</creatorcontrib><creatorcontrib>Wang, Zhehong</creatorcontrib><creatorcontrib>Li, Ziyun</creatorcontrib><creatorcontrib>Wang, Guanru</creatorcontrib><creatorcontrib>Kim, Hun-Seok</creatorcontrib><creatorcontrib>Blaauw, David</creatorcontrib><creatorcontrib>Sylvester, Dennis</creatorcontrib><title>A 22nm 3.5TOPS/W Flexible Micro-Robotic Vision SoC with 2MB eMRAM for Fully-on-Chip Intelligence</title><title>2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits)</title><addtitle>VLSI Technology and Circuits</addtitle><description>We present a highly flexible micro-robotic vision SoC featuring a hybrid Processing Element (PE) for efficient processing of both Convolutional Neural Network (CNN) and non-CNN vision tasks with 2MB embedded MRAM for retentive fully-on-chip weight storage. Fabricated in 22nm, the design achieves 0.22nJ/pix for Harris corner detection (a non-CNN vision task) and 3.5TOPS/W (INT16) for CNN, a 60% efficiency improvement over state-of-the-art NVM-based NN ASICs.</description><subject>Convolutional neural networks</subject><subject>Corner detection</subject><subject>Task analysis</subject><subject>Very large scale integration</subject><issn>2158-9682</issn><isbn>9781665497725</isbn><isbn>1665497726</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2022</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNot0E1PwjAcgPFqYiIin8BLv8BmX9a3Iy6iJCwYQDzi2v0HNaUl24zy7T3I6bn9Dg9COSU5pcQ8bhfr-QbcIaaQ9uc6NqXvCqmkyRlhLDeaE16QKzQxSlMpRWGUYuIajRgVOjNSs1t01_dfhDAiuB6hzylmLB4xz8Vm-bZ-_MCzAL_eBsCVd13KVsmmwTu89b1PEa9TiX_8cMCsesJQraYVblOHZ98hnLMUs_LgT3geBwjB7yE6uEc3bR16mFw6Ru-z5035mi2WL_Nyusg8pXrIrJEF19o5STnUWhCjaq0UIbKVDa8dtVAIy5VuC1OAAkUsF1YpLVvXkIbxMXr4dz0A7E6dP9bdeXcZwv8AnM1Y7Q</recordid><startdate>20220612</startdate><enddate>20220612</enddate><creator>Zhang, Qirui</creator><creator>An, Hyochan</creator><creator>Fan, Zichen</creator><creator>Wang, Zhehong</creator><creator>Li, Ziyun</creator><creator>Wang, Guanru</creator><creator>Kim, Hun-Seok</creator><creator>Blaauw, David</creator><creator>Sylvester, Dennis</creator><general>IEEE</general><scope>6IE</scope><scope>6IH</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIO</scope></search><sort><creationdate>20220612</creationdate><title>A 22nm 3.5TOPS/W Flexible Micro-Robotic Vision SoC with 2MB eMRAM for Fully-on-Chip Intelligence</title><author>Zhang, Qirui ; An, Hyochan ; Fan, Zichen ; Wang, Zhehong ; Li, Ziyun ; Wang, Guanru ; Kim, Hun-Seok ; Blaauw, David ; Sylvester, Dennis</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i118t-b964388cc613ea85097a877006f6d3ac1be45b378f494e7e70b35b7786fcd0d23</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Convolutional neural networks</topic><topic>Corner detection</topic><topic>Task analysis</topic><topic>Very large scale integration</topic><toplevel>online_resources</toplevel><creatorcontrib>Zhang, Qirui</creatorcontrib><creatorcontrib>An, Hyochan</creatorcontrib><creatorcontrib>Fan, Zichen</creatorcontrib><creatorcontrib>Wang, Zhehong</creatorcontrib><creatorcontrib>Li, Ziyun</creatorcontrib><creatorcontrib>Wang, Guanru</creatorcontrib><creatorcontrib>Kim, Hun-Seok</creatorcontrib><creatorcontrib>Blaauw, David</creatorcontrib><creatorcontrib>Sylvester, Dennis</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan (POP) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Electronic Library Online</collection><collection>IEEE Proceedings Order Plans (POP) 1998-present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Zhang, Qirui</au><au>An, Hyochan</au><au>Fan, Zichen</au><au>Wang, Zhehong</au><au>Li, Ziyun</au><au>Wang, Guanru</au><au>Kim, Hun-Seok</au><au>Blaauw, David</au><au>Sylvester, Dennis</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>A 22nm 3.5TOPS/W Flexible Micro-Robotic Vision SoC with 2MB eMRAM for Fully-on-Chip Intelligence</atitle><btitle>2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits)</btitle><stitle>VLSI Technology and Circuits</stitle><date>2022-06-12</date><risdate>2022</risdate><spage>72</spage><epage>73</epage><pages>72-73</pages><eissn>2158-9682</eissn><eisbn>9781665497725</eisbn><eisbn>1665497726</eisbn><abstract>We present a highly flexible micro-robotic vision SoC featuring a hybrid Processing Element (PE) for efficient processing of both Convolutional Neural Network (CNN) and non-CNN vision tasks with 2MB embedded MRAM for retentive fully-on-chip weight storage. Fabricated in 22nm, the design achieves 0.22nJ/pix for Harris corner detection (a non-CNN vision task) and 3.5TOPS/W (INT16) for CNN, a 60% efficiency improvement over state-of-the-art NVM-based NN ASICs.</abstract><pub>IEEE</pub><doi>10.1109/VLSITechnologyandCir46769.2022.9830340</doi><tpages>2</tpages></addata></record>
fulltext fulltext_linktorsrc
identifier EISSN: 2158-9682
ispartof 2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits), 2022, p.72-73
issn 2158-9682
language eng
recordid cdi_ieee_primary_9830340
source IEEE Xplore All Conference Series
subjects Convolutional neural networks
Corner detection
Task analysis
Very large scale integration
title A 22nm 3.5TOPS/W Flexible Micro-Robotic Vision SoC with 2MB eMRAM for Fully-on-Chip Intelligence
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-20T05%3A23%3A07IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_CHZPO&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=A%2022nm%203.5TOPS/W%20Flexible%20Micro-Robotic%20Vision%20SoC%20with%202MB%20eMRAM%20for%20Fully-on-Chip%20Intelligence&rft.btitle=2022%20IEEE%20Symposium%20on%20VLSI%20Technology%20and%20Circuits%20(VLSI%20Technology%20and%20Circuits)&rft.au=Zhang,%20Qirui&rft.date=2022-06-12&rft.spage=72&rft.epage=73&rft.pages=72-73&rft.eissn=2158-9682&rft_id=info:doi/10.1109/VLSITechnologyandCir46769.2022.9830340&rft.eisbn=9781665497725&rft.eisbn_list=1665497726&rft_dat=%3Cieee_CHZPO%3E9830340%3C/ieee_CHZPO%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-i118t-b964388cc613ea85097a877006f6d3ac1be45b378f494e7e70b35b7786fcd0d23%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=9830340&rfr_iscdi=true