Improving the Efficiency of Power Management Techniques by Using Bayesian Classification

Hwisung Jung and Massoud Pedram
University of Southern California


This paper presents a supervised learning based dynamic power management (DPM) framework for a multicore processor, where a power manager (PM) learns to predict the system performance state from some readily available input features (such as the state of service queue occupancy and the task arrival rate) and then uses this predicted state to look up the optimal power management action from a pre-computed policy lookup table. The key rationale for utilizing supervised learning in the form of a Bayesian classifier in the DPM context is to reduce the overhead of the PM which has to monitor the workload of the system and continually issue voltage-frequency mode transition commands to each processor core in the system. Experimental results demonstrate that the proposed Bayesian Classification based DPM technique ensures robust system-wide energy savings under rapidly and widely varying workloads.