Actionable Program Analyses for Improving Software Performance

Abstract

AbstractNowadays, we have greater expectations of software than ever before. This is followed by the constant pressure to run the same program on smaller and cheaper machines. To meet this demand, the application’s performance has become an essential concern in software development. Unfortunately, many applications still suffer from performance issues: coding or design errors that lead to performance degradation. However, finding performance issues is a challenging task: there is limited knowledge on how performance issues are discovered and fixed in practice, and current profilers report only where resources are spent, but not where resources are wasted. In this chapter, we investigate actionable performance analyses that help developers optimize their software by applying relatively simple code changes. We focus on optimizations that are effective, exploitable, recurring, and out-of-reach for compilers. These properties suggest that proposed optimizations lead to significant performance improvement, that they are easy to understand and apply, applicable across multiple projects, and that the compilers cannot guarantee that these optimizations always preserve the original program semantics. We implement our actionable analyses in practical tools and demonstrate their potential in improving software performance by applying relatively simple code optimizations.