High-level software energy macro-modeling

This paper presents an efficient and accurate high-level software energy estimation methodology using the concept of characterization-based macro-modeling. In characterization-based macro-modeling, a function or sub-routine is characterized using an accurate lower-level energy model of the target processor, to construct a macro-model that relates the energy consumed in the function under consideration to various parameters that can be easily observed or calculated from a high-level programming language description. The constructed macro-models eliminate the need for significantly slower instruction-level interpretation or hardware simulation that is required in conventional approaches to software energy estimation. We present two different approaches to macro-modeling for embedded software that offer distinct efficiency-accuracy characteristics: (i) complexity-based macro-modeling, where the variables that determine the algorithmic complexity of the function under consideration are used as macro-modeling parameters, and (ii) profiling-based macro-modeling, where internal profiling statistics for the functions are used as parameters in the energy macro-models. We have experimentally validated our software energy macro-modeling techniques on a wide range of embedded software routines and two different target processor architectures. Our experiments demonstrate that high-level macro-models constructed using the proposed techniques are able to estimate the energy consumption to within 95% accuracy on the average, while commanding speedups of one to five orders-of-magnitude over current instruction-level and architectural energy estimation techniques.