Process and User Driven Dynamic Voltage and Frequency Scaling

We propose and evaluate two new and independently-applicable techniques, processdriven voltage scaling (PDVS) and user-driven frequency scaling (UDFS), for improved power management on processors that support Dynamic Voltage and Frequency Scaling (DVFS), e.g, those used in current laptop and desktop computers. In PDVS, a CPUcustomized profile is derived offline that encodes the minimum voltage needed to achieve stability at each combination of CPU frequency and temperature. On a typical processor, PDVS reduces the voltage far below the worst-case minimum operating voltages given in datasheets. UDFS, on the other hand, dynamically adapts CPU frequency to the individual user and the workload through a simple user feedback mechanism, unlike currently-used DVFS methods which rely only on CPU utilization. Our UDFS algorithms dramatically reduce typical operating frequencies while maintaining performance at satisfactory levels for each user. We evaluated our techniques independently and together through user studies conducted on a modern Pentium M laptop running Windows applications. Our studies include both single task and multitasking scenarios. We measure the overall system power and temperature reduction caused by our methods. Combining PDVS and the best UDFS scheme reduces measured system power by 49.9% (27.8% PDVS, 22.1% UDFS), averaged across all our users and applications, compared to the Windows XP DVFS scheme. The average temperature of the CPU is decreased by 13.2 C on average. Using user trace-driven simulation to evaluate the CPU only, we find average CPU dynamic power savings of 57.3% (32.4% PDVS, 24.9% UDFS), with a maximum of reduction 83.4%. In a multitasking environment, the same UDFS+PDVS technique reduces the CPU dynamic power by 75.7% on average. Effort sponsored by the National Science Foundation under Grants ANI-0093221, ANI-0301108, EIA0224449, and CCF-0541337, and by the Department of Energy through grant FG02-05ER25691.