Energy-aware real-time scheduling on Heterogeneous Multi-Processor

Heterogeneous Multi-Processor has been steadily gaining importance because of its potential energy efficiency compared to homogeneous multi-processor architectures. Many challenges for scheduling algorithms on heterogeneous multi-process are created during the shift from homogeneous multi-processor to heterogeneous architectures, especially for real-time applications. In this paper, we introduce an optimal algorithm, using energy-aware real-time scheduling, for multiple tasks on heterogeneous multi-core processors. We first propose an energy modelling from both dynamic and static energy perspectives, considering the parameters both of software and hardware levels. After that we propose a real-time scheduling policy of Optimal Job to a Fast Processor First (OJFPF) using the parameter of energy achieved from energy modelling, the length of tasks as well as the local priority to assign the priority to each task. Our simulation results show that OJFPF algorithm using the energy modelling significantly reduce the overall energy consumption and improve the performance.