J. Kim, Shinhaeng Kang, Sukhan Lee, Hyeonsu Kim, Woongjae Song, Yuhwan Ro, Seungwon Lee, David Wang, Hyunsung Shin, BengSeng Phuah, Jihyun Choi, J. So, Yeon-Gon Cho, Joonho Song, J. Choi, Jeonghyeon Cho, Kyomin Sohn, Y. Sohn, Kwang-il Park, N. Kim
{"title":"Aquabolt-XL: Samsung HBM2-PIM with in-memory processing for ML accelerators and beyond","authors":"J. Kim, Shinhaeng Kang, Sukhan Lee, Hyeonsu Kim, Woongjae Song, Yuhwan Ro, Seungwon Lee, David Wang, Hyunsung Shin, BengSeng Phuah, Jihyun Choi, J. So, Yeon-Gon Cho, Joonho Song, J. Choi, Jeonghyeon Cho, Kyomin Sohn, Y. Sohn, Kwang-il Park, N. Kim","doi":"10.1109/HCS52781.2021.9567191","DOIUrl":null,"url":null,"abstract":"Using PIM to overcome memory bottleneck • Although various bandwidth increase methods have been proposed, it is physically impossible to achieve a breakthrough increase. - Limited by # of PCB wires, # of CPU ball, and thermal constraints • PIM has been proposed to improve performance of bandwidth-intensive workloads and improve energy efficiency by reducing computing-memory data movement.","PeriodicalId":246531,"journal":{"name":"2021 IEEE Hot Chips 33 Symposium (HCS)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"17","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 IEEE Hot Chips 33 Symposium (HCS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/HCS52781.2021.9567191","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 17
Abstract
Using PIM to overcome memory bottleneck • Although various bandwidth increase methods have been proposed, it is physically impossible to achieve a breakthrough increase. - Limited by # of PCB wires, # of CPU ball, and thermal constraints • PIM has been proposed to improve performance of bandwidth-intensive workloads and improve energy efficiency by reducing computing-memory data movement.