A new DVS prediction algorithm for ubiquitous devices based on Cycle Mode Matching

To take advantage of the full potential of ubiquitous devices, we will need systems which minimize power consumption. In this paper, a new time-interval-based Dynamic Voltage Scaling (DVS) prediction algorithm, Cycle Mode Matching (CMM) is proposed through reasonable analysis of the prediction parts of main current DVS algorithms. The simulation results show that the accuracy of DVS prediction is vastly improved especially for ubiquitous devices without the perceptible impact of system performance.