Semi-Federated Scheduling of Parallel Real-Time Tasks on Multiprocessors

Xu Jiang,Nan Guan,Xiang Long,Wang Yi

DOI: https://doi.org/10.48550/arXiv.1705.03245

2017-05-09

Abstract:Federated scheduling is a promising approach to schedule parallel real-time tasks on multi-cores, where each heavy task exclusively executes on a number of dedicated processors, while light tasks are treated as sequential sporadic tasks and share the remaining processors. However, federated scheduling suffers resource waste since a heavy task with processing capacity requirement $x + \epsilon$ (where $x$ is an integer and $0 < \epsilon < 1$) needs $x + 1$ dedicated processors. In the extreme case, almost half of the processing capacity is wasted. In this paper we propose the semi-federate scheduling approach, which only grants $x$ dedicated processors to a heavy task with processing capacity requirement $x + \epsilon$, and schedules the remaining $\epsilon$ part together with light tasks on shared processors. Experiments with randomly generated task sets show the semi-federated scheduling approach significantly outperforms not only federated scheduling, but also all existing approaches for scheduling parallel real-time tasks on multi-cores.

Distributed, Parallel, and Cluster Computing