Title:
Semantics-Aware Performance Optimization

Abstract:
Over the past decade, the pervasive use of advanced object-oriented languages like Java and the increasing complexity of tasks accomplished by software have led to the proliferation of large framework-intensive applications. These applications commonly suffer from performance problems that are hard to detect and optimize away. As many of these problems are caused by programmers' mistakes and inappropriate design choices, they cannot be easily removed by existing optimization techniques, which are often unaware of the semantics of the target program. In this talk, we argue that future optimization research should focus on bringing high-level developer insight into optimization instead of developing sophisticated techniques that are still semantics-agnostic. I will talk about two specific semantics-aware optimization techniques we have developed, one for memory leak detection and the other for automated container optimization.