Scenario-based run-time adaptive MPSoC systems

详细信息    查看全文

• 作者：Wei Quan ; ^a ; ^b ; ^{w.quan@uva.nl" class="auth_mail" title="E-mail the corresponding author} ; ^{quanwei02@gmail.com" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Andy D. Pimentel^a ; ^{a.d.pimentel@uva.nl" class="auth_mail" title="E-mail the corresponding author}Author Vitae
• 关键词：Embedded systems ; Multiprocessor system-on-chip (MPSoC) ; Design space exploration ; Task mapping ; Adaptive systems ; Reconfiguration
• 刊名：Journal of Systems Architecture
• 出版年：2016
• 出版时间：January 2016
• 年：2016
• 卷：62
• 期：Complete
• 页码：12-23
• 全文大小：2163 K

文摘

The ever-increasing performance demand of modern embedded applications drives the development of multi-processor system-on-chip (MPSoC) systems in the embedded domain. Today’s MPSoC-based products increasingly have to deal with multiple application execution scenarios which may change dynamically at run time. To improve the system performance, a state-of-the-art solution is to dynamically adapting the allocation of system resources at run time for each execution scenario based on pre-determined resource schemes that have been optimized at design time. However, such approaches will not work well for MPSoC systems that have a large number of execution scenarios and/or frequent run-time variations in execution scenario behavior. In this work, we therefore propose a scalable run-time self-adaptive framework for MPSoC systems that addresses these problems, thereby considerably improving the system efficiency.