A genetic algorithm to improve linux kernel performance on resource-constrained devices

Abstract

As computers become increasingly mobile, users demand more functionality, longer battery-life, and better performance from mobile devices. In response, chipset fabricators are focusing on elegant architectures to provide solutions that are both low-power and high-performance. Since these architectures rely on unique x86 extensions rather than fast clock speeds and large caches, careful thought must be placed into effective optimization strategies for not only user applications, but also the kernel itself, as the typical default optimizations used by modern compilers do not often take advantage of these specialized features. Focusing on the Intel Diamondville platform, this paper presents a genetic algorithm that evolves the compiler flags needed to build a Linux kernel that exhibits reduced response times.