No abstract

Background: Performance bugs can lead to severe issues regarding computation efficiency, power consumption, and user experience. Locating these bugs is a difficult task because developers have to judge for every costly operation whether runtime is consumed necessarily or unnecessarily. Objective: We wanted to investigate how developers, when locating performance bugs, navigate through the code, understand the program, and communicate the detected issues. Method: We performed a qualitative user study observing twelve developers trying to fix documented performance bugs in two open source projects. The developers worked with a profiling and analysis tool that visually depicts runtime information in a list representation and embedded into the source code view. Results: We identified typical navigation strategies developers used for pinpointing the bug, for instance, following method calls based on runtime consumption. The integration of visualization and code helped developers to understand the bug. Sketches visualizing data structures and algorithms turned out to be valuable for externalizing and communicating the comprehension process for complex bugs. Conclusion: Fixing a performance bug is a code comprehension and navigation problem. Flexible navigation features based on executed methods and a close integration of source code and performance information support the process.

No abstract

Due to non-deterministic behavior and thread interleaving of concurrent programs, the debugging of concurrency and performance issues is a rather difficult and often tedious task. In this paper, we present an approach that combines statistical profiling, clustering and visualization to facilitate this task. We implemented our approach in a tool which is integrated as a plugin into a widely used IDE. First, we introduce our approach with details on the profiling and clustering strategy that produce runtime metrics and clusters of threads for source-code artifacts at different levels of abstraction (class and method) and the entire program. Next, we explain the design of our visualization which represents the clusters in situ, i.e., embedded in the program text next to the related source-code artifact in the source-code editor. More detailed information is available in separate windows that also allow the user to configure thread filters interactively. In a demonstration study, we illustrate the usefulness of the tool for understanding and fixing performance and concurrency issues. Finally, we report on first formative results from a usability test and consequently implemented tool improvements. Graphical abstract