Data Compression Techniques for Improving Memory Hierarchy Performance

点击翻译按钮获取中文摘要

Abstract:This project is aimed at developing data compression techniques forsubstantially reducing the data memory footprint of memory intensiveapplications. We will study the behavior of a wide range of applicationsto identify characteristics of data held in data structures by theseprograms that can be exploited in carrying out data compression. Wewill develop a suite of data compression transformations that will beimplemented by the compiler. Our goal is to develop data transformationsthat can be applied to partially compressible data structures, that is,data structures in which all the data is not compressible. This approachwill not only result in transformations that are widely applicable, butalso lead to techniques for applying them that are efficient. In particular,instead of relying on complex compile-time analysis for proving theirapplicability, we will be able to use simple value profiling techniquesto identify data structures that are mostly compressible and then applythe transformations to them. We will develop data compression extensions(DCX) to a RISC-style instruction set for efficiently manipulatingcompressed data. In addition, we will also address other low level codegeneration issues, such as impact of transformations on register pressureand instruction cache behavior, which can have a significant impact onperformance.