Memory and compiler optimizations for low-power and -energy

Abstract

Embedded systems become more and more widespread, especially autonomous ones, and clearly tend to be ubiquitous. In such systems, low-power and low-energy usage get ever more crucial. Furthermore, these issues also become paramount in (massively) multi-processors systems, either in one machine or more widely in a grid. The various problems faced pertain to autonomy, power supply possibilities, thermal dissipation, or even sheer energy cost. Although it has since long been studied in harware, energy optimization is more recent in software. In this paper, we thus aim at raising awareness to low-power and low-energy issues in the language and compilation community. We thus broadly but briefly survey techniques and solutions to this energy issue, focusing on a few specific aspects in the context of compiler optimizations and memory management.