Power Management in Real Time Embedded Systems through Online and Adaptive Interplay of DPM and DVFS Policies

• Main Authors: Bhatti, K, Belleudy, C, Auguin, M
• Format: Conference Proceeding
• Language: English
• Subjects: Dynamic Voltage & Frequency Scaling; Energy consumption; Energy-efficient Scheduling; Machine Learning; Power demand; Processor scheduling; Program processors; Real time systems; Runtime; Time frequency analysis
• DOI: 10.1109/EUC.2010.35

This paper considers the problem of power/energy minimization for periodic real-time tasks that are scheduled over multiprocessor platforms that have dynamic power management (DPM) and dynamic voltage & frequency scaling (DVFS) capabilities. Early research reports that while both DPM and DVFS policies perform well individually for a specific set of conditions, they often outperform each other under different workload and/or architecture configuration. Thus, no single policy fits perfectly all operating conditions. Instead of designing new policies for specific operating conditions, this paper proposes a generic power management scheme, called the Hybrid Power Management (HyPowMan) scheme. This scheme takes a set of well-known existing (DPM and DVFS) policies, each of which performs well for a given set of conditions, and adapts at runtime to the best-performing policy for any given workload. We performed experiments with state-of the-art DPM and DVFS techniques and results show that HyPowMan scheme adapts well to the changing workload and always achieves overall energy savings comparable to the best-performing policy at any point in time.