Abstract
In preemptive real-time systems, scheduling analyses need - in addition to the worst-case execution time - the context-switch cost. In case of preemption, the preempted and the preempting task may interfere on the cache memory. These interferences lead to additional reloads in the preempted task. The delay due to these reloads is referred to as the cache-related preemption delay (CRPD). The CRPD constitutes a large part of the context-switch cost. In this article, we focus on the computation of upper bounds on the CRPD based on the concepts of useful cache blocks (UCBs) and evicting cache blocks (ECBs). We explain how these concepts can be used to bound the CRPD in case of direct-mapped caches. Then we consider set-associative caches with LRU, FIFO, and PLRU replacement. We show potential pitfalls when using UCBs and ECBs to bound the CRPD in case of LRU and demonstrate that neither UCBs nor ECBs can be used to bound the CRPD in case of FIFO and PLRU. Finally, we sketch a new approach to circumvent these limitations by using the concept of relative competitiveness.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.