Browsers are the operating systems of the Web. They support a vast universe of applications written in a modern garbage-collected programming language. Browsers expose a rich platform API mostly implemented in C++. Browsers are also consumer software with low switching costs in an intensely competitive market. Thus in addition to standard requirements such as maximizing throughput and minimizing latency, browsers have to consider issues like-when the user closes a window while watching Task Manager, they want to see memory usage go down. Browsers have to compete to minimize memory usage even for poorly written applications. In this talk I will elucidate these requirements and describe how Firefox and other browsers address them. I will pay particular attention to issues that we don't know how to solve, and that could benefit from research attention.

Instrumenting programs with code to monitor runtime behavior is a common technique for profiling and debugging. In practice, instrumentation is either inserted manually by programmers, or automatically by specialized tools that monitor particular properties. We propose Program Trace Query Language (PTQL), a language based on relational queries over program traces, in which programmers can write expressive, declarative queries about program behavior. We also describe our compiler, Partiqle. Given a PTQL query and a Java program, Partiqle instruments the program to execute the query on-line. We apply several PTQL queries to a set of benchmark programs, including the Apache Tomcat Web server. Our queries reveal significant performance bugs in the jack SpecJVM98 benchmark, in Tomcat, and in the IBM Java class library, as well as some correct though uncomfortably subtle code in the Xerces XML parser. We present performance measurements demonstrating that our prototype system has usable performance.

We present Object Equality Profiling (OEP), a new technique for helping programmers discover optimization opportunities in programs. OEP discovers opportunities for replacing a set of equivalent object instances with a single representative object. Such a set represents an opportunity for automatically or manually applying optimizations such as hash consing, heap compression, lazy allocation, object caching, invariant hoisting, and more. To evaluate OEP, we implemented a tool to help programmers reduce the memory usage of Java programs. Our tool performs a dynamic analysis that records all the objects created during a particular program run. The tool partitions the objects into equivalence classes, and uses collected timing information to determine when elements of an equivalence class could have been safely collapsed into a single representative object without affecting the behavior of that program run. We report the results of applying this tool to benchmarks, including two widely used Web application servers. Many benchmarks exhibit significant amounts of object equivalence, and in most benchmarks our profiler identifies optimization opportunities clustered around a small number of allocation sites. We present a case study of using our profiler to find simple manual optimizations that reduce the average space used by live objects in two SpecJVM benchmarks by 47% and 38% respectively.

We present a new method for dynamically detecting potential data races in multithreaded programs. Our method improves on the state of the art in accuracy, in usability, and in overhead. We improve accuracy by combining previously known race detection techniques -- lockset-based detection and happens-before-based detection -- to obtain fewer false positives than lockset-based detection alone. We enhance usability by reporting more information about detected races than any previous dynamic detector. We reduce overhead compared to previous detectors -- particularly for large applications such as Web application servers -- by not relying on happens-before detection alone, by introducing a new optimization to discard redundant information, and by using a two phase approach to identify error-prone program points and then focus instrumentation on those points. We justify our claims by presenting the results of applying our tool to a range of Java programs, including the widely-used Web application servers Resin and Apache Tomcat. Our paper also presents a formalization of locksetbased and happens-before-based approaches in a common framework, allowing us to prove a folk theorem that happens-before detection reports fewer false positives than lockset-based detection (but can report more false negatives), and to prove that key optimizations are correct.

We present a novel approach to dynamic datarace detection for multithreaded object-oriented programs. Past techniques for on-the-fly datarace detection either sacrificed precision for performance, leading to many false positive datarace reports, or maintained precision but incurred significant overheads in the range of 3x to 30x. In contrast, our approach results in very few false positives and runtime overhead in the 13% to 42% range, making it both efficient and precise. This performance improvement is the result of a unique combination of complementary static and dynamic optimization techniques.

No abstract available.