Practical models for energy-efficient prefetching in mobile embedded systems

Energy efficiency is the most important concern in mobile embedded system design. The conventional wisdom is that there is a tradeoff between energy efficiency and high-performance techniques, such as prefetching. Thus to reduce energy consumption and save chip area, hardware prefetchers are not implemented in most existing embedded mobile systems. However, modern embedded mobile systems have become increasingly powerful and show a great deal of demand for applying high-performance techniques, such as hardware prefetching, to accelerate applications. In this paper, we study whether it would be beneficial to implement hardware prefetchers in embedded mobile systems. We first demonstrate that: contrary to the conventional wisdom, as technology advances (e.g. from 90 nm to 32 nm), prefetching starts to become energy-efficient while improving performance. Then, we introduce a general analytical model to identify the conditions for prefetching techniques to achieve energy efficiency. Furthermore, we also introduce a series of models to evaluate the energy efficiency of the prefetcher when specific prefetching features are employed. By using these models, system designers can easily and accurately evaluate the energy efficiency of their designs and make decisions on the deployment of hardware prefetchers.