Optimal task allocation and scheduling for power saving in multiprocessor systems

Abstract

The wide spread use of multiprocessor system demands for energy saving. Energy issues being the most critical one need prime attention and prevention. This paper presents methods for power consumption reduction in multiprocessor systems by optimally allocating the jobs and scheduling them. The job allocation in multiprocessor systems is very crucial and a novel allocation algorithm for reducing the number of processors is proposed in this work. Scheduling decisions are vital for the system to meet both temporal and functional requirements and is perfect when methods for energy consumption reduction are incorporated into it. The current work aims to reduce power consumption of an embedded processor by Dynamic Voltage Frequency Scaling (DVFS) technique. The simulation results show a considerable amount of reduction in processor utilization and power consumption.