A multi-granularity energy profiling approach and a quantitative study of a Web browser

Abstract

While energy-efficiency considerations become a critical issue in embedded-system designs, little work is done for energy-profiling at different granularity levels. This paper aims at the proposing of a hybrid hardware-software-based profiling solution with different levels of abstraction. Compared to the existing work, the proposed solution could provide richer information on energy profiling with the help of execution-path tracking and exclusion of selected functions/classes/modules. We address technical issues in the implementation of the proposed profiling solution. A quantitative energy-profiling study over a well-known Web browser based on an object-oriented design, Konqueror, is then presented to provide a feasibility study of the proposed solution and insights in the design of an energy-efficient Web browser.