Is Semi-Partitioned Scheduling Practical?

Abstract

Semi-partitioned schedulers are-in theory-a particularly promising category of multiprocessor real-time scheduling algorithms. Unfortunately, issues pertaining to their implementation have not been investigated in detail, so their practical viability remains unclear. In this paper, the practical merit of three EDF-based semi-partitioned algorithms is assessed via an experimental comparison based on real-time schedulability under consideration of real, measured overheads. The presented results indicate that semi-partitioning is indeed a sound and practical idea. However, several problematic design choices are identified as well. These shortcomings and other implementation concerns are discussed in detail.

Resource Sharing under Multiprocessor Semi-partitioned Scheduling

“…With big overhead values, generally the platforms become more unschedulable under MRPS. This is an important factor which makes the MRPS algorithm suitable for multiprocessor platforms with tasks having small working set size (WSS) [3].…”

“…Once the job executes its critical section at destination processor, it will migrate back to its dedicated processor (source processor). As a result of migrations, the subtasks will incur overhead mainly due to cache-related migration delay [3]. This overhead is caused due to additional cache misses that a job incurs when resuming execution after a migration.…”

Simulation of Real-time Multiprocessor Scheduling with Overheads

“…Following these observations, recent research has emerged and new scheduling algorithms appeared which aim to reduce the overheads by bounding the amount of preemptions (Bastoni et al, 2011;Nelissen et al, 2012). Also, a few studies have shown that avoiding co-scheduling tasks that heavily use a shared cache can reduce the overall execution time (Fedorova et al, 2006;Anderson et al, 2006).…”

“…In that case, there is a single queue of ready tasks and a single scheduler for all the processors. Finally, as a compromise that aims to alleviate limitations of partitioned (limited achievable processor utilization) and global (non-negligible overheads) algorithms, hybrid policies such as semi-partitioned and clustered scheduling have been proposed more recently (Bastoni et al, 2011).…”

rt-muse: measuring real-time characteristics of execution platforms

“…Despite it being a complete workload, it does not provide any data analysis. Bastoni et al (2011) studied semi-partitioned schedulers in depth, evaluating the scheduling overhead, the cache-induced delays and other metrics with LITMUS RT (Calandrino et al 2006). While both rt-muse and (Bastoni et al 2011) use traces of events to analyze the behavior of schedulers, LITMUS RT aims at simplifying the implementation of complex scheduling policies, while rt-muse aims at closing the gap between theory and implementation by exposing the real-time characteristics of already implemented schedulers.…”