Energy Efficient Greedy Scheduling of Tasks for DVFS Enabled Heterogeneous Multicore Processors

Nowadays the demand for high performance computing systems like servers, increasing since these systems, process many real-time applications with the help of heterogeneous processors. However, in multicore systems power consumption is always a concern. In most of the processor based systems dynamic voltage and frequency scaling is a common method adopted to save energy. In processing many heterogeneous tasks, it is appropriate to propose a combined energy efficient scheduling using DVFS. In this article, we introduce a task scheduling model along with DVFS method for high performance computing system. The proposed greedy scheduler assigns the tasks to the suitable processor core based on their energy profile and probability value of the incoming task. The simulation results show that the proposed scheduling method bounds for the objective of minimal energy consumption of the target system.