Task assignment with energy efficiency considerations for non-DVS heterogeneous multiprocessor systems

Abstract

Currently, the multiprocessor architecture platform becomes more and more popular to the applications with higher computational capacity for meeting the fast-grown demand of applications and increasing the performance of whole system. Managing the consumption of resources has become an important issue in many different settings. The energyefficient scheduling for multiprocessors becomes one of important research issues. In this paper we study the scheduling problem of a heterogeneous non-DVSmultiprocessor platform with a task set. The processors have different characteristics of power consumption. We propose an off-line taskto- processor assignment algorithm, the Best-Fit Decreasing Physical Power Consumption (BDPC) algorithm to derive a feasible task assignment with the minimal energy consumption and has the time complexity of O ( N ( logN + M )), where N and M are the numbers of tasks and processor types, respectively. A series of experiments were conducted to evaluate the proposed algorithm. The experimental results demonstrate that the performance of the proposed BDPC algorithm is better than the compared algorithms.