{"title":"Invited talk: Memory: The center of the universe","authors":"R. Murphy","doi":"10.1109/WMED.2013.6544499","DOIUrl":null,"url":null,"abstract":"Summary form only given. The memory system is typically the primary driver behind system performance, which has been acknowledged as the von Neumann bottleneck since the invention of the modern electronic computer in the 1940s. Tremendous unsustainable investments have been made on the processor side to attempt to get around the problem. However, the end of Dennard Scaling in 2003 meant that the memory wall must finally be addressed to support the transition to simpler, multicore architectures. Consequently, computer architects must consider the fundamental system-level tradeoffs, rather than myopically focusing on CPU core architecture. This talk will describe the first decade of the transition into the new era of computer architecture, its impact on application performance, and what we as a community can do about it.","PeriodicalId":134493,"journal":{"name":"2013 IEEE Workshop on Microelectronics and Electron Devices (WMED)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2013-04-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2013 IEEE Workshop on Microelectronics and Electron Devices (WMED)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/WMED.2013.6544499","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Summary form only given. The memory system is typically the primary driver behind system performance, which has been acknowledged as the von Neumann bottleneck since the invention of the modern electronic computer in the 1940s. Tremendous unsustainable investments have been made on the processor side to attempt to get around the problem. However, the end of Dennard Scaling in 2003 meant that the memory wall must finally be addressed to support the transition to simpler, multicore architectures. Consequently, computer architects must consider the fundamental system-level tradeoffs, rather than myopically focusing on CPU core architecture. This talk will describe the first decade of the transition into the new era of computer architecture, its impact on application performance, and what we as a community can do about it.