Introduction
Users
Documentation
Developers
Availability
Download


Applications


paraprof viewer


paraprof viewer


racy viewer

vampir viewer


Examples

pdgssvx
psgesv

TAU

Introduction

TAU (Tuning and Analysis Utilities) is a set of tools for analyzing the performance of C, C++, Fortran and Java programs. It is normally used as follows:

  1. Instrument the program by inserting TAU macros into the program (this can be done automatically).
  2. Run the program. Files containing information about the program performance are automatically generated.
  3. View the results with TAU's pprof, the TAU visualizer paraprof (or racy), or a third-party visualizer (such as VAMPIR)
TAU collects much more information than what is available through prof or gprof, the standard Unix utilities. Also available through TAU are:
  • Per-process, per-thread and per-host information (supports pthreads).
  • Inclusive and exclusive function times.
  • Profiling groups that allow you to organize data collection.
  • Access to hardware counters on some systems.
  • Per-class and per-instance information.
  • Separate data for each template instantiation.
  • Start/stop timers for profiling arbitrary sections of code.
  • Support for collection of statistics on user-defined events.
TAU is designed so that when you turn off profiling (by disabling TAU macros) there is no overhead. Also, TAU uses the Program Database Toolkit (PDT), which is a tool infrastructure for static and dynamic analysis of object-oriented software. PDT provides access to the high-level interface of source code for analysis tools and applications.

Users

TAU can be used by anyone writing a C, C++, Fortran or Java application and who wants to understand where the performance bottlenecks are. The original users of TAU were associated with the ASCI projects at Los Alamos National Laboratory, which sponsored a good part of TAU's development. Two other users are the developers of Blitz++, a high performance C++ library that performs vector and array operations, and the ZPL group at the University of Washington. TAU is particularly useful to these users because of its support for template classes and expression templates. The images in the left panel show performance data generated by TAU. Click on the icons to display the full image (a new window will open).

Documentation

The TAU website contains a User's Guide, a list of publications, a list of Frequently Asked Questions, and also a self-running demo.

Developers

TAU was developed at the University of Oregon and at the Advanced Computing Laboratory at Los Alamos National Laboratory. Its principal developers were Sameer Shende, Allen Malony, Janice Cuny, Kathie Lindlan, Peter Beckman and Steve Karmesin.




Tools News Home