Open Research Position: Execution-Driven Performance Analysis for Distributed/Parallel Systems

Open Research Position: Execution-Driven Performance Analysis for Distributed/Parallel Systems

Post by Thomas Fahringe » Sat, 01 Apr 2000 04:00:00



Dear Colleague,

I have a position for a doct*student or a post-doc available at
the University of Vienna. I'd be grateful if you would pass
the adverti*t below on to any suitably-qualified person
who might be interested in it.

Please accept my sincere apologies if you receive multiple copies
of this job-offer.

Best Regards
Thomas Fahringer

---------------------------------------------------------------------
The Institute for Software Science at the University of Vienna
is offering within a long-term research project a

    ***********************************************************
    ***  Research Position in Execution-Driven Performance  ***
    ***    Analysis for Distributed and Parallel Systems    ***
    ***********************************************************

Applicants should have knowledge in one or more of the
following areas:

        + distributed and parallel systems
        + performance measurement, monitoring, and tracing
        + on-line and post-execution performance analysis
        + performance visualization
        + programming skills
            + distributed programming (Java)
            + parallel programming (OpenMP, HPF, MPI, threads, ...)
        + databases and expert systems

Position: Research position for a doct*student or a post-doc.

Starting Date: April 10, 2000

Duration: 3 years

Project Summary:

   The position is offered as part of a long-term research project
   about performance-oriented application development for distributed
   and parallel systems.

   The main task of this position requires to develop a performance
   analysis system that searches for performance problems in
   heterogeneous, object-oriented multi-threaded distributed and
   parallel applications exploiting both data and task parallelism.
   These applications are executed on clusters of SMPs or on
   heterogeneous workstation networks. Various instrumentation systems
   are used to obtain raw performance data while executing a program.
   Performance data may be stored in a data repository (database or
   expert system) for post-execution analysis. The system to be developed
   tries to find performance problems based on collected and computed
   performance data, and information about the input program provided by
   a compiler. Performance problems will be associated with the input program.
   Based on detected performance problems further decisions may be taken, e.g.
   refinement of instrumentation, more detailed search for performance
   problems, application and system changes to improve performance, etc.

For more information, please contact


=======================================================================
Thomas Fahringer, Ph.D.             Tel: (office): +43 1 310 56 08 - 86
Associate Professor                 Tel: (sec):    +43 1 310 56 08 - 71  
University of Vienna                Fax: +43 1 310 56 08 - 88

Liechtensteinstr. 22                WWW: http://www.veryComputer.com/~tf
A-1090 Vienna, Austria

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

p.s. Note that there is another open research position at the Institute
for Software Science, University of Vienna. For more details see below:
Make sure that you indicate in your reply whether you apply for the research
position about execution-driven performance analysis or about performance
prediction.

The Institute for Software Science at the University of Vienna is offering
within a long-term research project a

    ********************************************************
    ***  Research Position in Performance Prediction of  ***
    ***     Distributed and Parallel Applications        ***
    ********************************************************

Applicants should have knowledge in one or more of the
following areas:

        + distributed and parallel systems
        + programming skills
            + distributed programming (Java)
            + parallel programming (OpenMP, HPF, MPI, threads, ...)
        + performance modeling (analytical + symbolic + simulation)
        + simulation tools (discrete event simulation)

Position: Research position for a doct*student or a post-doc.

Starting Date: April 10, 2000

Duration: 3 years

Project Summary:

   The position is offered as part of a long-term research project
   about performance-oriented application development for parallel
   and distributed systems.

   The main task of this position requires to develop a performance
   prediction tool for heterogeneous, object-oriented multi-threaded
   parallel and distributed applications exploiting both data and
   task parallelism. These applications are executed on clusters
   of SMPs or on heterogeneous workstation networks.
   Analytical performance prediction will be used to determine
   parameterized cost functions for small components of a
   parallel/distributed application. Parameters in cost functions
   reflect the problem sizes of an application and the machine sizes
   of a target architecture. Simulation will be used to model
   highly dynamic behavior of applications and architectures, in
   particular, data exchange, synchronization, thread context-switches,
   etc. The performance prediction tool computes various performance
   parameters including estimated execution and communication times,
   synchronization overhead, memory locality, load balance, etc.

For more information, please contact


--

======================================================================
Thomas Fahringer, Ph.D.              Tel: (office): +43 1 310 56 08 - 86
Associate Professor                  Tel: (sec):    +43 1 310 56 08 - 71  
University of Vienna                 Fax: +43 1 310 56 08 - 88

Liechtensteinstr. 22                 WWW: http://www.veryComputer.com/
A-1090 Vienna, Austria