Evaluation of software energy consumption on microprocessors

Abstract

In the area of wireless communications, energy consumption is the key design consideration. Significant effort has been placed in optimizing hardware for energy efficiency, while relatively less emphasis has been placed on software energy reduction. For overall energy efficiency reduction of system energy consumption in both hardware and software must be addressed. One goal of this research is to evaluate the factors that affect software energy efficiency and identify techniques that can be employed to produce energy optimal software. In order to present a strong argument, two state-of-the-art low power processors were used for evaluation: the Intel StrongARM SA-1100 and the next generation Intel Xscale processor. A key step in analyzing the performance of software is to perform a comprehensive tabulation of the energy consumption per instruction, while taking into account the different modes of operation. This leads into a comprehensive energy profiling for the instruction set of the processors of interest. With information on the energy consumption per instruction, we can evaluate the feasibility of energy efficient programming and use the results to gain greater insight into the power consumption of the two processors under consideration. Benchmark programs will be tested on both processors to illustrate the effectiveness of the energy profiling results. The next goal is to look at the leakage current and current consumed during idle modes of the processors and how that impacts the overall picture of energy consumption. Thus energy consumption will be explored for the two processors from both a dynamic and static energy consumption perspective.