This paper presents novel sampling-based techniques for collecting statistical profiles of register contents, data values, and other information associated with instructions, such as memory latencies. Values of interest are sampled in response to periodic interrupts. The resulting value profiles can be analyzed by programmers and optimizers to improve the performance of production uniprocessor and multiprocessor systems.Our value sampling system extends the DCPI continuous profiling infrastructure, and inherits many of its desirable properties: our value profiler has low overhead (approximately 10% slowdown); it profiles all the code in the system, including the operating system kernel; and it operates transparently, without requiring any modifications to the profiled code.

Article Continuous profiling: where have all the cycles gone? Share on Authors: Jennifer M. Anderson View Profile , Lance M. Berc View Profile , Jeffrey Dean View Profile , Sanjay Ghemawat View Profile , Monika R. Henzinger View Profile , Shun-Tak A. Leung View Profile , Richard L. Sites View Profile , Mark T. Vandevoorde View Profile , Carl A. Waldspurger View Profile , William E. Weihl View Profile Authors Info & Claims SOSP '97: Proceedings of the sixteenth ACM symposium on Operating systems principlesOctober 1997 Pages 1–14https://doi.org/10.1145/268998.266637Published:01 October 1997 175citation1,209DownloadsMetricsTotal Citations175Total Downloads1,209Last 12 Months17Last 6 weeks3 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

This article describes the Digital Continuous Profiling Infrastructure, a sampling-based profiling system designed to run continuously on production systems. The system supports multiprocessors, works on unmodified executables, and collects profiles for entire systems, including user programs, shared libraries, and the operating system kernel. Samples are collected at a high rate (over 5200 samples/sec. per 333MHz processor), yet with low overhead (1–3% slowdown for most workloads). Analysis tools supplied with the profiling system use the sample data to produce a precise and accurate accounting, down to the level of pipeline stalls incurred by individual instructions, of where time is bring spent. When instructions incur stalls, the tools identify possible reasons, such as cache misses, branch mispredictions, and functional unit contention. The fine-grained instruction-level analysis guides users and automated optimizers to the causes of performance problems and provides important insights for fixing them.

Many parallel applications are designed to conceal parallelism from the user. We investigate a different approach where the user controls many tasks running in parallel. The idea is to let a user accomplish his goal more quickly by trying competing alternatives in parallel (or-parallelism) and by working on subgoals in parallel (and-parallelism). To help the user manage a large number of parallel tasks, the application must provide features to generate many tasks easily, to summarize the state of all tasks, to broadcast commands to related tasks, and to abort tasks that are no longer needed. A parallel interface to an application thus becomes crucial to enhance the user's productivity. We demonstrate this approach using DLP, a parallel, distributed version of the Larch Prover, an interactive theorem prover. DLP supports explicit parallelism and runs on a network of workstations. Users control DLP through a multi-window interface on a bit-map color-display. Many theorem proving problems that would otherwise take considerable user effort to solve have been done with relative ease using DLP.

Managing tradeoffs between program structure and program efficiency is one of the most difficult problems facing software engineers. Decomposing programs into abstractions simplifies the construction and maintenance of software and results in fewer errors. However, the introduction of these abstractions often introduces significant inefficiencies.This paper describes a strategy for eliminating many of these inefficiencies. It is based upon providing alternative implementations of the same abstraction, and using information contained in formal specifications to allow a compiler to choose the appropriate one. The strategy has been implemented in a prototype compiler that incorporates theorem proving technology.