Compiler Techniques for Code Compression

University Of Arizona, Tucson AZ

Investigators

Abstract

The goal of this project is to develop techniques to reduce the memory footprint of executable code, so as to allow more and more sophisticated applications to be executed on limited-memory devices, such as hand-held computers, personal digital assistants, and embedded processors. Recent years have seen the incorporation of computers and computational devices into many aspects of our everyday lives. In many cases, the amount of memory available for such processors is limited by considerations such as space, weight, and power consumption. At the same time, there is a desire to run more and more sophisticated applications on such processors. Since an application that occupies more memory than is available on such a processor will not be able to run on that processor, it is desirable to develop techniques to reduce the memory footprint of applications. Moreover, it is necessary that the compressed applications remain executable, since for the application domains under consideration it is not feasible to decompress the executable in order to execute it. This project investigates the construction of tools and techniques for code compression in a manner that preserves executability.