Parameter sensitivity analysis of the Energy/Frequency Convexity Rule for application processors

Abstract

Both theoretical and experimental evidence are presented in this work in order to validate the existence of an Energy/Frequency Convexity Rule, which relates energy consumption and microprocessor frequency for nanometer-scale microprocessors. Data gathered during several month-long experimental acquisition campaigns, supported by several independent publications, suggest that energy consumed is indeed depending on the microprocessor's clock frequency, and, more interestingly, the curve exhibits a clear minimum over the processor's frequency range. An analytical model for this behavior is presented and motivated, which fits well with the experimental data. A parameter sensitivity analysis shows how parameters affect the energy minimum in the clock frequency space. The conditions are discussed under which this convexity rule can be exploited, and when other methods are more effective, with the aim of improving the computer system's energy management efficiency. We show that the power requirements of the computer system, besides the microprocessor, and the overhead affect the location of the energy minimum the most. The sensitivity analysis of the Energy/Frequency Convexity Rule puts forward a number of simple guidelines especially for by low-power systems, such as battery-powered and embedded systems, and less likely by high-performance computer systems.

Experimental energy profiling of energy-critical embedded applications

“…It is obvious that the energy consumption curves are convex, each having an optimal frequency point (f opt ) where the energy consumption is minimized. This is consistent with previous experiments revealing the existence of the Energy/Frequency Convexity Rule for compute-intensive applications running on an Exynos-based platform in a Samsung Galaxy S2 phone [2], [6].…”

“…Other setups often only provide access to platform-level energy consumption, which includes the energy consumed by the peripheral devices, power management unit, GPU, memory, et cetera. A more complete state of the art can be found in [6].…”

“…We revalidate the Energy/Frequency Convexity Rule using much more precise methodology and tools than in our previous work [2], [6], which is of interest on its own. The level of accuracy achieved in this work will allow us in the future to precisely correlate the segments of a software program to its energy trace.…”

Efficient mobile code offloading in heterogeneous wireless networks

“…The energy consumption of mobile devices can be reduced mainly in two ways: one way is code offloading, which, in some situations, may cause more energy waste, and hence, another way is to use big.LITTLE technology on the mobile device itself. The computation energy increased with the growth of CPU frequency when the task was executed on both the little core and the big core, and the energy consumption of the big core was higher than that of the little core 18 …”

Scheduling-based power capping in high performance computing systems

“…Since the HPC community widely recognizes the need to reduce power consumption in supercomputers, several research avenues have been explored for this purpose, in particular power capping solutions. Many techniques have been proposed to bound the power consumption of HPC machines, ranging from Dynamic Voltage and Frequency Scaling (DVFS) [8,9,10,11,12], energy proportional systems [13], overprovisioning [14], turning off idle resources [15], exploiting components variability [16] , optimizing the placements of jobs and tasks in order to reduce communication costs [17]. In the rest of this Section 2.1 we are going to discuss power capping methods found in the literature.…”