商用48层和96层3D-NAND闪存中的SLC和MLC内存近似搜索解决方案

P. Tseng, Tian-Cig Bo, Yu-Hsuan Lin, Yu-Chao Lin, Jhe-Yi Liao, F. Lee, Yu-Yu Lin, Ming-Hsiu Lee, K. Hsieh, Keh-Chung Wang, Chih-Yuan Lu
{"title":"商用48层和96层3D-NAND闪存中的SLC和MLC内存近似搜索解决方案","authors":"P. Tseng, Tian-Cig Bo, Yu-Hsuan Lin, Yu-Chao Lin, Jhe-Yi Liao, F. Lee, Yu-Yu Lin, Ming-Hsiu Lee, K. Hsieh, Keh-Chung Wang, Chih-Yuan Lu","doi":"10.1109/IMW56887.2023.10145964","DOIUrl":null,"url":null,"abstract":"We proposed a novel in-memory-approximate-search (IMAS) method from single-level cell (SLC) to multi-level cell (MLC) based on our commercial 48-layer (48L) and 96-layer (96L) 3D-NAND flash technology. The method with word-line input as search word provides ultra-high parallel searching capability with the database stored in high-density 3D NAND-flash IMAS chip(s). The input search word can be compared with $128\\mathrm{~K}$ data words in just one read cycle, where the Hamming distance (HD) computation (for SLC approach) or similarity computation (for MLC approach) can be carried out within the memory array. Latest 96L 3D-NAND IMAS chip with CMOS under array (CuA) technology provides small chip size, long search/data word, excellent output resolution, and wide matching criteria for the tolerance on sensing current variation. Application on Memory-Augmented Neural Network (MANN) was demonstrated on VGGface2 dataset, with the comparisons between the 48L and 96L SLC 3D-NAND IMAS technologies.","PeriodicalId":153429,"journal":{"name":"2023 IEEE International Memory Workshop (IMW)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"SLC and MLC In-Memory-Approximate-Search Solutions in Commercial 48-layer and 96-layer 3D-NAND Flash Memories\",\"authors\":\"P. Tseng, Tian-Cig Bo, Yu-Hsuan Lin, Yu-Chao Lin, Jhe-Yi Liao, F. Lee, Yu-Yu Lin, Ming-Hsiu Lee, K. Hsieh, Keh-Chung Wang, Chih-Yuan Lu\",\"doi\":\"10.1109/IMW56887.2023.10145964\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We proposed a novel in-memory-approximate-search (IMAS) method from single-level cell (SLC) to multi-level cell (MLC) based on our commercial 48-layer (48L) and 96-layer (96L) 3D-NAND flash technology. The method with word-line input as search word provides ultra-high parallel searching capability with the database stored in high-density 3D NAND-flash IMAS chip(s). The input search word can be compared with $128\\\\mathrm{~K}$ data words in just one read cycle, where the Hamming distance (HD) computation (for SLC approach) or similarity computation (for MLC approach) can be carried out within the memory array. Latest 96L 3D-NAND IMAS chip with CMOS under array (CuA) technology provides small chip size, long search/data word, excellent output resolution, and wide matching criteria for the tolerance on sensing current variation. Application on Memory-Augmented Neural Network (MANN) was demonstrated on VGGface2 dataset, with the comparisons between the 48L and 96L SLC 3D-NAND IMAS technologies.\",\"PeriodicalId\":153429,\"journal\":{\"name\":\"2023 IEEE International Memory Workshop (IMW)\",\"volume\":\"59 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2023-05-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2023 IEEE International Memory Workshop (IMW)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/IMW56887.2023.10145964\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2023 IEEE International Memory Workshop (IMW)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IMW56887.2023.10145964","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

基于我们的商用48层(48L)和96层(96L) 3D-NAND闪存技术,我们提出了一种新的内存近似搜索(IMAS)方法,从单级单元(SLC)到多级单元(MLC)。该方法以字行输入作为检索词,将数据库存储在高密度3D NAND-flash IMAS芯片中,提供了超高的并行检索能力。输入的搜索词可以在一个读取周期内与$128\ mathm {~K}$数据词进行比较,其中可以在内存数组内进行汉明距离(HD)计算(用于SLC方法)或相似性计算(用于MLC方法)。最新的96L 3D-NAND IMAS芯片采用阵列下CMOS (CuA)技术,芯片尺寸小,搜索/数据字长,输出分辨率高,感应电流变化公差匹配标准宽。在VGGface2数据集上演示了记忆增强神经网络(MANN)的应用,并对48L和96L SLC 3D-NAND IMAS技术进行了比较。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
SLC and MLC In-Memory-Approximate-Search Solutions in Commercial 48-layer and 96-layer 3D-NAND Flash Memories
We proposed a novel in-memory-approximate-search (IMAS) method from single-level cell (SLC) to multi-level cell (MLC) based on our commercial 48-layer (48L) and 96-layer (96L) 3D-NAND flash technology. The method with word-line input as search word provides ultra-high parallel searching capability with the database stored in high-density 3D NAND-flash IMAS chip(s). The input search word can be compared with $128\mathrm{~K}$ data words in just one read cycle, where the Hamming distance (HD) computation (for SLC approach) or similarity computation (for MLC approach) can be carried out within the memory array. Latest 96L 3D-NAND IMAS chip with CMOS under array (CuA) technology provides small chip size, long search/data word, excellent output resolution, and wide matching criteria for the tolerance on sensing current variation. Application on Memory-Augmented Neural Network (MANN) was demonstrated on VGGface2 dataset, with the comparisons between the 48L and 96L SLC 3D-NAND IMAS technologies.
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