文章基本信息

标题：Analysis of Memory-Contention in Heterogeneous COTS MPSoCs
本地全文：下载
作者：Mohamed Hassan ; Rodolfo Pellizzoni
期刊名称：LIPIcs : Leibniz International Proceedings in Informatics
电子版ISSN：1868-8969
出版年度：2020
卷号：165
页码：23:1-23:24
DOI：10.4230/LIPIcs.ECRTS.2020.23
出版社：Schloss Dagstuhl -- Leibniz-Zentrum fuer Informatik
摘要：Multiple-Processors Systems-on-Chip (MPSoCs) provide an appealing platform to execute Mixed Criticality Systems (MCS) with both time-sensitive critical tasks and performance-oriented non-critical tasks. Their heterogeneity with a variety of processing elements can address the conflicting requirements of those tasks. Nonetheless, the complex (and hence hard-to-analyze) architecture of Commercial-Off-The-Shelf (COTS) MPSoCs presents a challenge encumbering their adoption for MCS. In this paper, we propose a framework to analyze the memory contention in COTS MPSoCs and provide safe and tight bounds to the delays suffered by any critical task due to this contention. Unlike existing analyses, our solution is based on two main novel approaches. 1) It conducts a hybrid analysis that blends both request-level and task-level analyses into the same framework. 2) It leverages available knowledge about the types of memory requests of the task under analysis as well as contending tasks; specifically, we consider information that is already obtainable by applying existing static analysis tools to each task in isolation. Thanks to these novel techniques, our comparisons with the state-of-the art approaches show that the proposed analysis provides the tightest bounds across all evaluated access scenarios.
关键词：DRAM; Memory; COTS; Multi-core; Real-Time; Embedded Systems; Analysis