{"title":"Memory Access Pattern Profiling for Streaming Applications Based on MATLAB Models","authors":"T. Goldbrunner, Thomas Wild, A. Herkersdorf","doi":"10.1109/PATMOS.2018.8464151","DOIUrl":null,"url":null,"abstract":"In order to cope with the problems caused by the Data Deluge, for application-specific architectures it is important to design the system in such a way that the processing and memory subsystems are geared to each other. Especially for systems that need to fulfill high data throughput demands, there is a high risk, that otherwise there will be high degradation of the overall system performance caused by memory accesses. However especially for application-specific designs, it is possible to create a tailored memory subsystem by profiling the applications memory access patterns beforehand and then incorporating the information during the system design. For this purpose we present a data access pattern profiling framework, which can provide detailed information about an applications memory access patterns based on a MATLAB model of the application.","PeriodicalId":234100,"journal":{"name":"2018 28th International Symposium on Power and Timing Modeling, Optimization and Simulation (PATMOS)","volume":"281 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2018 28th International Symposium on Power and Timing Modeling, Optimization and Simulation (PATMOS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/PATMOS.2018.8464151","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
In order to cope with the problems caused by the Data Deluge, for application-specific architectures it is important to design the system in such a way that the processing and memory subsystems are geared to each other. Especially for systems that need to fulfill high data throughput demands, there is a high risk, that otherwise there will be high degradation of the overall system performance caused by memory accesses. However especially for application-specific designs, it is possible to create a tailored memory subsystem by profiling the applications memory access patterns beforehand and then incorporating the information during the system design. For this purpose we present a data access pattern profiling framework, which can provide detailed information about an applications memory access patterns based on a MATLAB model of the application.