Newsgroups: comp.parallel
From: Mark SAWLEY <sawley@dgm.epfl.ch>
Subject: COSMASE/COMETT Parallel Computation course
Organization: Ecole Polytechnique Federale de Lausanne
Date: Mon, 23 Jan 1995 12:54:03 GMT
Message-ID: <3ftkk7$ipa@info.epfl.ch>


                   COSMASE/COMETT Advanced Course on


                 P A R A L L E L   C O M P U T A T I O N
                 --------------------------------------- 


    THEME:  THE EFFICIENT USE OF HIGH PERFORMANCE PARALLEL COMPUTERS


                 March 20-24 1995, Lausanne, Switzerland



                         FIRST ANNOUNCEMENT


COSMASE is a series of specialized training courses for participants
from both industry and academic institutions, organized within the
framework of the European COMETT Programme. The next course on
Parallel Computation in the COSMASE series will be held in Lausanne.

The aim of this practical course is to provide a broad overview of
the essential components of present-day high performance parallel
computing. The courses are of week-long duration, with lectures in
the mornings followed by practical sessions in the afternoons.
They are designed as specialized training courses for particpants
from both industry and academic institutions.

Emphasis is placed in this course on both hardware and software aspects
of direct importance for efficient implementation of scientific
applications. Of central interest are distributed-memory parallel
systems (both networked and integrated) that use cache-based RISC
processors, and programming models using both explicit (message passing)
and implicit (data parallel, virtual shared memory) inter-processor
communication. Code portability through the use of emerging standards
(e.g., HPF, MPI) is also addressed.

The course is intended for engineers who wish to exploit the
performance offered by high performance parallel computers. The course
will enable participants to obtain practical experience on a range of
the latest-generation parallel computer systems (e.g., Cray T3D,
Intel Paragon, IBM SP2, PCI/Meiko CS2, MasPar MP-1, workstation clusters),
and thus provide an essential base for current and potential users.


 PROGRAMME
 ---------

Monday  20 March 1995

08h00 - 08h20   Registration
08h20 - 08h30   Welcome and course overview
08h30 - 10h15   Introduction to high-performance parallel computing
10h45 - 12h30   Cache-based RISC processors: architecture and
                programming techniques
12h30 - 14h00   Lunch
14h00 - 18h00   Hands-on training session: Design of code for RISC
                processors (HP workstation and/or Cray T3D)


Tuesday  21 March 1995

08h30 - 10h15   Introduction to networked parallel systems:
                message passing programming model
10h45 - 12h30   Message passing using PVM
12h30 - 14h00   Lunch
14h00 - 18h00   Hands-on training session: Parallelization using PVM
                (workstation clusters and/or Cray T3D)


Wednesday  22 March 1995

08h30 - 10h15   Introduction to distributed-memory parallel computers
10h45 - 12h30   Data parallel programming model
12h30 - 14h00   Lunch
14h00 - 18h00   Hands-on training session: Use of the data parallel
                programming model (MasPar MP-1 and/or Cray T3D)


Thursday  23 March 1995

08h30 - 10h15   Shared memory programming model: Cray Research
10h45 - 12h30   Parallelization tools and evolution of programming models
12h30 - 14h00   Lunch
14h00 - 18h00   Hands-on training session: Use of the shared memory
                model on the Cray T3D


Friday  24 March 1995

08h30 - 13h00   Experiences with parallelizing scientific applications:
                - Parallel computing in aerodynamics: a collaborative
                  effort between industry and research centres
                - Parallel computation of compressible turbomachinery flows
                - Industrial flow simulations on parallel systems
                - Parallelization of RADIOSS, an industrial finite element
                  crash simulation software
13h00 - 14h00   Lunch
14h00 - 16h15   Hands-on training session: Use of programming model and
                parallel system of the participant's choice
                (Cray T3D, Intel Paragon, IBM SP2, PCI/Meiko CS2,
                MasPar MP-1, workstation cluster)


 LECTURERS
 ---------

U. Block           GENIAS Software GmbH, Neutraubling
L.J. Clarke        Edinburgh Parallel Computing Centre
M. Dayde           ENSEEIHT-INPT and CERFACS, Toulouse
L. Drummond        CERFACS, Toulouse
R. Greimel         AVL-List GmbH, Graz
C.-W. Hustad       Ecole Polytechnique Federale de Lausanne
D. Nicolopoulous   MECALOG, Sophia-Antipolis
M. Roche           Cray Research, Lausanne
M.L. Sawley        Ecole Polytechnique Federale de Lausanne
L. Tourrette       Aerospatiale Avions, Toulouse

------------------------------------------------------------------------

For more detailed information

          contact:  Dr. Mark L. Sawley
                    IMHEF / EPFL
                    ME-Ecublens
                    CH-1015 Lausanne,  Switzerland
                    Tel: +41 21 693 3531
                    Fax: +41 21 693 3646
                    email: sawley@dgm.epfl.ch

   World Wide Web:  http://imhefwww.epfl.ch/COSMASE



