Motivating complexity understanding by profiling energy usage

Abstract

Computer science and software engineering students are typically taught to evaluate resource use in terms of time complexity. Developers use asymptotic analysis to compare algorithms by calculating how time grows as a function of input size. However, two factors have limited traditional models of complexity as pedagogical tools. First, modern systems are so fast that even relatively inefficient algorithms can quickly process large sets of data. Second, analysis is not universally engaging; only some students care about efficiency for the sake of efficiency. Our project proposes using measurements of energy consumption and the concomitant environmental impact to better engage students with efficiency and its implications.