Enhancement and assessment of a code-analysis-based energy estimation framework

Energy estimation of applications helps developers greening the smartphone- and Internet-of-Things-based devices. Traditional energy estimation schemes consider smartphone component's power measurement or code analysis methods for energy estimation of applications. The existing code analysis method considers the energy cost of software operations to minimize the energy estimation overhead of dynamic estimation methods. However, it overlooked cache storage analysis and overheads associated with it due to concurrent program execution at runtime. As a result, the performance of estimation tools is affected. To handle these issues, this study put forward an enhanced static-code-analysis-based lightweight energy estimation (SA-LEE) framework that has considered overheads associated with the application runtime execution environment, cache storage analysis, and the application inactivity period for energy estimation of applications. The experiments revealed that the SA-LEE model has minimized the estimation time and the energy overhead by 98% and 97%, respectively. Also, the accuracy is observed to be 82-88%.