CORDIS
EU research results

CORDIS

English EN

Dynamic re-allocation of meshes for parallel finite element applications

Objective

The objectives of the DRAMA project are:

- To develop parallel mesh re-allocation algorithms, and a corresponding subroutine library, for unstructured Finite Element codes (including industrial Structural Mechanics codes) based on the DRAMA cost model.
- To demonstrate and validate the dynamic mesh re-allocation approaches with leading industrial codes (PAM-CRASH/PAM-STAMP, FORGE3).
- To develop a cost model of unstructured finite element applications which includes:
- dynamically changing computational and communications requirements;
- re-allocation costs.
- To enable efficient 3-D parallel adaptive solvers
- To enable fully parallel mesh generation as a by-product via exploitation of the parallel re-partitioning of adaptively generated meshes.

DRAMA will develop a library of routines for dynamic mesh re-allocation which will be integrated with MPI-based parallel unstructured Finite-Element applications - the DRAMA library is parallel and call-able at runtime from the applications. The applications will provide their computation and communication requirements, via the DRAMA cost function, and will receive the specification of the new partition and information on a data re-allocation strategy for it. The innovation of the DRAMA library lies in the ability to handle both computation and communication requirements resulting from the complexity occurring in structural mechanics codes and in the inclusion of library features which will take account of the machine architecture: hierarchical or clustered computing will be accounted for both the re-allocation costs and the re-allocation strategy.

The DRAMA project will result in an application programming technology which will allow complex industrial simulation codes to progress beyond the use of static mesh partitioning parallelisation in order to provide their users with full and efficient code functionality on HPC systems.

Parallel repartitioning is recognised as a key component which will allow the extensive use, in the near future, of large to massively parallel architectures for scientific, industrial computing. It is the only way to guarantee the scalability, and thus the efficient resource exploitation, of advanced simulation codes.
Leaflet | Map data © OpenStreetMap contributors, Credit: EC-GISCO, © EuroGeographics for the administrative boundaries

Coordinator

Engineering Systems Intl

Address

Rue Saarinen 20
94150 Rungis

France

Participants (2)

Sort alphabetically

Expand all

Association pour la Recherche et le Developpement des Methodes et Processus Industriels

France

KATHOLIEKE UNIVERSITEIT LEUVEN

Belgium

Project information

Grant agreement ID: FP4_24953

  • Start date

    1 October 1997

  • End date

    30 September 1999

Funded under:

FP4-ESPRIT 4

  • Overall budget:

    € 1 830 000

  • EU contribution

    € 1 012 500

Coordinated by:

Engineering Systems Intl

France