Virtual Institute – High Productivity Supercomputing

Euro-Par'13 tutorial: Tools for High Productivity Supercomputing (RWTH, Aachen, Germany)

Date

Monday 26th August 2013.

Presenters

  • Brian Wylie, Jülich Supercomputing Centre
  • Martin Schulz, Lawrence Livermore National Laboratory

and featuring additional tools demonstrations by:

  • Joachim Protze, RWTH Aachen University
  • Ronny Tschüter, TU Dresden
  • Isaías Comprés Ureña, TU München

Location

The tutorial takes place as part of the Euro-Par'13 conference in Aachen, Germany, in a classroom of the RWTH computing centre (RZ) which provides computers for participants. These can be used to follow the presentations and demonstrations using the VI-HPS Linux ISO where many of the VI-HPS tools are installed for use.

Outline

Developers are challenged to improve the reliability, performance and scalability of their applications on computer systems with increasingly large numbers of processors and cores, requiring ever more parallelism and efficiency with many processes and threads of execution. Use of appropriate tools can greatly assist productivity of both developers and computer systems, and is the founding motivation for the Virtual Institute High Productivity Supercomputing (VI-HPS). This full-day tutorial will give an overview and introduction to tools covering execution monitoring, correctness checking, and performance analysis of parallel applications at large to extreme scale. Primary focus is presenting and demonstrating tools provided on the VI-HPS Linux ISO for use in virtual environments that are mostly available as open-source and suited for a range of current HPC platforms and Linux clusters, such that attendees will know which tools to look for on their development and production platforms and how they can be applied to improve productivity.

Programme

09:00 Introduction
  • VI-HPS Overview & Linux ISO
  • Parallel application engineering & workflow
  • Extreme-scale case studies
  • VI-HPS productivity tools suite
  • 10:30 (break)
    11:00 Execution monitoring, checking & debugging
  • system/batchqueue monitoring (PTP/SysMon)
  • lightweight execution monitoring (LWM2)
  • portable performance counter access (PAPI)
  • MPI library profiling (mpiP)
  • MPI execution outlier detection (AutomaDeD)
  • MPI memory usage checking (memchecker)
  • MPI correctness checking (MUST)
  • lightweight stack trace debugging (STAT)
  • task dependency debugging (Temanejo)
  • 12:30
    (lunch)
    14:30 Integrated application execution profile & trace analysis
  • instrumentation & measurement (Score-P, Extrae)
  • profile analysis examination (Cube, ParaProf)
  • execution trace exploration (Vampir, Paraver)
  • automated trace analysis (Scalasca)
  • on-line automated analysis (Periscope)
  • 16:00 (break)
    16:30 Complementary tools & utiilities
  • parallel performance frameworks (O|SS, TAU)
  • cache usage analysis (callgrind/kcachegrind)
  • performance analysis data-mining (PerfExplorer)
  • parallel execution parametric studies (Dimemas)
  • parallel file I/O optimization (SIONlib)
  • process mapping generation/optimization (Rubik)
  • assembly code optimization (MAQAO)
  • etc.
  • Review & Discussion
  • 18:00 (adjourn)

    Registration

    Registration via the Euro-Par 2013 conference website is possible with or without including the conference itself.

    Contact

    Brian Wylie
    Jülich Supercomputing Centre
    Phone: +49 2461 61-6589
    Email: b.wylie@fz-juelich.de