Compiler-Directed Communication Energy Optimizations for Microsensor Networks

In this chapter, the authors focus on a wireless microsensor network environment that process array-intensive codes. They present the computation and communication energy models used in their work. The authors provide experimental data that show the effectiveness of their approach and propose a compiler-based communication optimization algorithm for microsensor networks. The energy behavior of applications running on these networks is largely dictated by the software support employed such as operating systems and compilers. A networked system of inexpensive and plentiful microsensors, with multiple sensor types, low-power embedded processors, and wireless communication and positioning ability offers a promising solution for many military and civil applications. The authors also provide an energy breakdown when different data decompositions and communication optimizations are used. They highlight the effectiveness of inter-nest communication optimization. Placing communication calls within the innermost loop can increase the energy spent in communication dramatically.