ReMaP--Regularity and Massive Parallelism
At its creation, the aim of the project was to bring new contributions and
software tools for massively parallel computing applications. Data-parallel
compilation is still an important part of the project but we are now
investigating the application and impact of heterogeneous
networks--libraries, virtual shared memory and communication protocols.
ReMaP researchers aim at designing methods and tools for parallelism at
every level: from high-speed network protocols to parallel-compilation
techniques, and from multi-threaded run-time environments to computing and
communication libraries for parallel applications. Within each of these
topics, we develop both theoretical work and software prototypes.
The LHPC is a common research laboratory operating on massively parallel
computers and high-performance computing. The main function of the LHPC is
to serve as a center of expertise in the field of massively parallel
computing by bringing together partners of academic and industrial projects
with worldwide ambitions.
Collaborators
Ecole Normale Superieure de Lyon (LIP), France
Innovative Computing Laboratory University of Tennessee, USA
Contact
Yves Robert
École Normale Supérieure de Lyon
Yves.Robert@lip.ens-lyon.fr
http://www.inria.fr/Equipes/REMAP-eng.html