Various Approaches for High Throughput and Energy Efficient Scheduling of Real-Time Tasks in Multicore Systems

Embedded real-time systems are gradually increasing to execute high-performance-oriented applications on multicore architecture. It is very necessary to have efficient task scheduling model in these systems so that majority of the tasks can be scheduled within their deadline to provide the needed throughput. We proposed a dynamic scheduling algorithm where random tasks are accommodated into a system that is already running a set of tasks, meeting the deadline of the tasks. Random tasks are generated at different time intervals with different periodicity and execution time. Energy consumption is one of the critical design issues in real-time systems especially in battery-operated systems. We further implemented an energy efficient real-time scheduling where voltages are adjusted dynamically to achieve optimum energy as well as throughput. We extended the scheduling work to heterogeneous multicore systems (HMS) architecture where real-time tasks are accommodated to the appropriate cores maintaining the task deadline. We demonstrated that the heterogeneous multicore scheduling model can be used as commercially available heterogeneous multicore systems. We verified the model using generated task sets and the results showed that our model performs excellently in all cases and significantly reduces the energy consumption of the system with respect to some popular and new scheduling schemes.