Skip to main content

An integrated programming environment for parallel sparse matrix solvers

Objective

The objectives are:

- to develop sparse matrix algorithms according to the needs of the industrial partners
- to implement portable prototypes of these solvers and to demonstrate these working on HPC systems under MPI
- to specify an open library interface for the PARASOL library
- to integrate the solvers and additional utilities into the PARASOL library
- to establish a continuous library support
- to use a state-of-the-art parallel programming environment
- to evaluate PARASOL solvers and utilities by industrial application developers

The PARASOL consortium consists of

- leading European research organisations with a well-known experience and track-record in the development of parallel solvers (CERFACS, GMD, ONERA, RAL, Univ. of Bergen)
- industrial code developers who define the requirements for PARASOL and will use its results (Det Norske Veritas, INPRO, MacNeal-Schwendler, Polyflow),
- the two leading European HPC software companies which will exploit the project results and will provide state-of-the-art programming development tools (GENIAS, PALLAS).
PARASOL will produce the following tangible results:

- specification of an open library interface for parallel solvers (which will be disseminated to other solver developers outside the project)
- solver prototypes based on MPI
- a prototype of the PARASOL library
- assessment and evaluation reports by industrial software developers
- technology transfer and commercial exploitation of the project results.
The PARASOL project will develop new parallel algorithms for the direct solution of sparse systems of equations and for their preconditioning. These algorithms will be implemented as portable prototypes and will be integrated into the PARASOL library. The library interface will be specified in analogy to the MAxx codes of the (sequential) Harwell Library. A public domain version will be made available.

In addition to the solvers, the PARASOL library will contain interfaces to domain decomposition algorithms, utilities for data sorting, symbolic factorisation, and mapping algorithms.

Industrial developers will play a key role in the project and will be actively involved in the specification of the library as well as in the evaluation of the prototype solvers.

Funding Scheme

CSC - Cost-sharing contracts

Coordinator

Gesellschaft für Parallele Anwendungen und Systeme mbH
Address
Hermühlheimer Straße 10
50321 Brühl
Germany

Participants (6)

CCLRC Rutherford Appleton Laboratory (RAL)
United Kingdom
Address
Chilton
OX11 0QX Didcot
Cerfacs
France
Address
Avenue Gustave Coriolis 42
31057 Toulouse
Fraunhofer Gesellschaft
Germany
Address
Schloss Birlinghoven
53754 Munich
Genias Software
Germany
Address
Erzgebirgstrasse 2
93073 Neutraubling
Office National D'etudes et de Recherches Aerospatiales
France
Address
Division Leclerc 29
92322 Chatillon
University of Bergen
Norway
Address
Thormoehlensgt. 55
5020 Bergen