Thesis (M.Sc.) - Cairo University - Faculty of Computers and Information- Department of Computer Science

Power-aware computing has caught the interest of researchers and users of all computing systems. In embedded systems and small devices, better management of energy translates into longer lasting and smaller batteries, which in turn implies smaller and lighter devices. In the cloud, distributed, and high-performance computing systems, better management of power translates into saving a significant amount of money and natural resources. Hence, energy saving has become a concern for designing the computing devices infrastructure and the systems and software that operates it. In most of the cases, the ability to improve energy efficiency is limited by performance requirements and physical constraints, which make reaching the balance between performance and energy consumption a difficult objective to achieve, for energy and power-aware applications. On the software side, the coding style along with the compiler choice have a significant impact on the application{u2019}s performance, power, and energy consumption.The way we write our code defines how the instructions will be executed, how the application performs, and how it consumes energy. When a program is executed on a computing device, it consumes energy based on how it uses the computing device{u2019}s resources. Each instruction inside the program contributes to the resources usage and to the total energy being consumed. Those instructions get generated by the used compiler, based on how the program{u2019}s code is written

Issued also as CD