A semi-partitioned real-time scheduling approach for periodic task systems on multicore platforms.

Semi-partitioned scheduling is regarded as a viable alternative to partitioned or global scheduling approaches. Advantage of semi-partitioned scheduling is two-folds: it has reduced runtime overhead compared to global scheduling, and improved schedulability and system utilization factor compared to partitioned scheduling. This paper proposes a new semi-partitioned scheduling algorithm for real-time periodic task systems over multicore platforms. Our proposed algorithm works in two phases. In the first phase, each task from a feasible application task set is statically assigned to a specific processor. If a task can not be partitioned on any processor in the platform, it qualifies as migrating task. In the second phase, processors are clustered together such that, per cluster, the unused fragmented computation power equivalent to at most one processor is available. We provide schedulability analysis and experimental evaluation to support our proposition. Moreover, simulation results show an average difference of 18-folds in the number of task preemptions and 10-folds in the number of task migrations compared to multiprocessor optimal scheduling algorithm PD.