ICSE: U: Identifiying caching opportunities by revisiting Object Equality Profiling

A memory profiler is a tool used by software developers to understand how an application uses the physical memory during its execution. Most programming environments are shipped with an accurate profiler. In case of an excessive memory consumption, a software developer may reduce the application memory footprint by examining a profiler report. Unfortunately, optimization opportunities are not apparent in such a report due to the limitations of traditional profilers. Optimizing an application requires deep and extended knowledge about the analyzed application. This paper presents and evaluates a new approach to identify redundant objects. Once their origin is precisely located, such objects may be shared to reduce the memory consumed by the application. This paper reproduces and evaluates a memory situation identified by Marinov and Callahan. Our approach improves their work to address the needs met by practitioners.