Abstract—In real-time task scheduling, semi-partitioning allows some tasks to be split into portions and each portion to be assigned to a different core. This improves the performance of system but by counting each portion as a separate task it increases effective number of tasks to be scheduled. This research suggests a semi-partitioning method and assigns each partition to a separate core to be scheduled by the well-known scheduler called Rate-Monotonic (RM). To assure nonconcurrent execution of portions of a task, there is no need to define release time for any portion. It is theoretically proven that with the proposed semi-partitioning and RM scheduling, all cores always run their tasks overrun-free. Besides, experimental results show that overall system utilization is noticeably boosted and also number of broken tasks is not higher than the best RMbased methods.

Keywords