Information processing at work: On a theory for experimental algorithm complexity

It is common experience to upgrade firmware of mobile devices and obtain longer battery life, living proof of how software affects power consumption of a device. Despite this empirical observation, there is a lack for models and methodologies correlating computations with power consumption [3-5]. In this paper we propose an experimental approach to computational complexity and a methodology for conducting measures which result independent of the underlying system running the algorithm/software to be tested. Early experimental results are presented and discussed, showing that our methodology is robust and can be used in many settings. We also introduce the foundations of a theory for experimental algorithm complexity, which mimics what is predicted by the classic theory of computational complexity (big-O or Theta notations), except for some notable exceptions that we highlight and comment. This theory is validated in many scenarios, by considering several architectures and algorithms.  Because of the relation between time complexity and energy consumption, we may suggest that our work measures the “information work”: namely, the energy required for performing information processing.