logo inria

News
INRIA
Scientific Research
Valorization and Transfert
Publications and Documentation
Working and Training at INRIA

Version française Directory Site map
 Advanced search and help

Same level information :

| Research teams | Cooperative Initiatives | International Scientific Collaboration | Researchers' News |

-----------------------
APACHE : Parallel algorithms and load sharing (project-team)
-----------------------

illustration projet About the project :
Project-Team site Activity report Videos and photos
Research reports
Theses

Theme : Num
Numerical systems

Grenoble - Rhône-Alpes research center


APACHE has been dissolved on 12/31/2004
It has been replaced by MESCAL
It has been replaced by MOAIS

Project-Team Presentation

ID Laboratory joint project-team with CNRS, INPG and l'UJF

APACHE is a parallel processing research project, hosted in the ID (Informatics and Distribution) laboratory of the Grenoble Institute for Computer Science and Applied Mathematics (IMAG), France. It is sponsored by CNRS, INPG, INRIA and UJF.

This project investigates a new approach to program parallel machines for compute-intensive applications. Our goal is to find a good trade-off between performance and portability, so that programs with various profiles can be executed efficiently on parallel machines with different resources and facilities.

The programming environment Athapascan is an answer to these issues: efficiency and portability. A multithreaded communication runtime system, Athapascan-0, has been designed and has proven to be adequate and efficient. A programming library prototype, Athapascan-1, is available. It is based on a multi-tasking programming model with data consistency and grants an automatic load balancing. Test applications are being developped on Athapascan: molecular dynamics, quantum chemistry, computer algebra, domain decomposition and discrete event simulation. Finally, a tracing tool allows debbuging, performance evaluation and visualization of Athapascan and its applications.

Research themes

Research themes

Research themes

  • Parallel algorithms and complexity.
  • Parallel Applications.
  • Programming models and runtime environments.
  • Observation of parallel programs, debugging and performance evaluation.

Scientific leader

Brigitte PLATEAU
+33 4 76 61 20 88
Brigitte.Plateau@inria.fr
--------------------------------
|  back to top    | homepage
© INRIA - updated : 02/18/2005 - webmaster@inria.fr