{"title":"基于NAND-SPIN MRAM的CNN片上训练近似计算","authors":"Zhengyi Hou, Luyao Shi, Bi Wang, Zhaohao Wang","doi":"10.1145/3565478.3572537","DOIUrl":null,"url":null,"abstract":"Approximate computation is a widely used method to accelerate CNN training. In this work, the stochastic switching mechanism of the NAND-SPIN MRAM is utilized to perform the approximate update and storage of the synaptic weight. By reducing the programming time of the NAND-SPIN MTJs from 3 ns to 1 ns, more than 67% speedup and nearly 70% energy saving have been achieved with less than 1% accuracy loss.","PeriodicalId":125590,"journal":{"name":"Proceedings of the 17th ACM International Symposium on Nanoscale Architectures","volume":"46 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Approximate computation based on NAND-SPIN MRAM for CNN on-chip training\",\"authors\":\"Zhengyi Hou, Luyao Shi, Bi Wang, Zhaohao Wang\",\"doi\":\"10.1145/3565478.3572537\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Approximate computation is a widely used method to accelerate CNN training. In this work, the stochastic switching mechanism of the NAND-SPIN MRAM is utilized to perform the approximate update and storage of the synaptic weight. By reducing the programming time of the NAND-SPIN MTJs from 3 ns to 1 ns, more than 67% speedup and nearly 70% energy saving have been achieved with less than 1% accuracy loss.\",\"PeriodicalId\":125590,\"journal\":{\"name\":\"Proceedings of the 17th ACM International Symposium on Nanoscale Architectures\",\"volume\":\"46 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-12-07\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the 17th ACM International Symposium on Nanoscale Architectures\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/3565478.3572537\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 17th ACM International Symposium on Nanoscale Architectures","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3565478.3572537","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Approximate computation based on NAND-SPIN MRAM for CNN on-chip training
Approximate computation is a widely used method to accelerate CNN training. In this work, the stochastic switching mechanism of the NAND-SPIN MRAM is utilized to perform the approximate update and storage of the synaptic weight. By reducing the programming time of the NAND-SPIN MTJs from 3 ns to 1 ns, more than 67% speedup and nearly 70% energy saving have been achieved with less than 1% accuracy loss.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
