Efficient Scheduling of Dependent Tasks in Many-Core Real-Time System Using a Hardware Scheduler

Abstract

This paper proposes an efficient hardware scheduler for scheduling dependent tasks in real-time many-core systems. The main idea behind the proposed scheduler is that the operating system selects tasks that can be scheduled with the Earliest Deadline First (EDF) algorithm and groups the related tasks according to their dependency. It then transfers the group information and scheduling specifications of each task to the scheduling hardware. The operating system uses the software EDF algorithm and manages the scheduling and assignment of the task to each processing core according to the dependencies of each task in the many-core system. The simulation results through comparison with previous work confirm the proposed hardware scheduler increases feasible tasks by 2.1 times, decreases miss tasks by 3.43 times, while also considering dependencies between tasks.