TAU Workshop, Harvard SEAS, 18 March 2014

Profiling Scientific Software Applications with the TAU Performance System.

Pierce Hall Room 213 (Brooks Room), 11:00am - 5:00pm EST

The TAU Performance System is a powerful toolkit for performance measurement and analysis of software written in a variety of languages and executing at all scales. This presentation will introduce TAU's profiling, tracing, and debugging support with focus on application to scientific software. We will cover performance evaluation of parallel programs written in Fortran, C, and C++, using parallel programming approaches like OpenMP and MPI. We will also demonstrate TAU's techniques for program instrumentation including automatic instrumentation of source code, compiler-based instrumentation, and library preloading. We will demonstrate how to gather performance data showing time spent in parallel regions, I/O, memory allocation, and hardware performance counters from PAPI. TAU's support for memory debugging and runtime fault isolation will also be demonstrated.

Using TAU on SEAS systems

For this workshop, use the TAU environment installed at /home/jlinford/TAU:

  • C-shell:
    source /home/jlinford/TAU/tau.csh
    
  • Bash:
    source /home/jlinford/TAU/tau.sh
    

Slides