Sites Inria

Version française

REMAP Research team

Regularity and massive parallel computing

  • Leader : Frédéric Desprez
  • Research center(s) : CRI Grenoble - Rhône-Alpes
  • Field : Numerical systems
  • Theme : Grids and high-performance computing

Team presentation

A few years ago the main targets of the project were algorithms and software tools for massively parallel computing. Data-parallel compilation is still an important part of the project. We are investigating new applications on heterogeneous networks (libraries, multi-threaded environments, communication protocols and their impact, algorithms for telecommunications).

ReMaP is a project of LIP (Laboratoire de l'Informatique du Parallélisme) jointly operated by CNRS, INRIA and École normale supérieure de Lyon. This project is located at Lyon.

Research themes

  • Tools for parallel linear algebra (algorithms for parallel dense and sparse linear algebra, design of numerical libraries on heterogeneous networks of processors).
  • Wireless networks protocols, approximation algorithms for scheduling and telecommunication problems.
  • Multi-threaded environments, network protocols for high performance PC clusters, and software for multimedia servers.
  • Automatic parallelization techniques.

International and industrial relations

  • Industrial collaborations with Sycomore Aérospatiale Matra, Alcatel, Microsoft, Sun Labs, HP Labs, Myricom, ProxyTV project.
  • International collaborations in academia (University of Tennessee, Knoxville (UT), University of California San Diego (UCSD), University of New Hampshire (UNH), Moscow, Berlin, and Bergen Universities).

Keywords: Massive parallelism Distributed-memory Networks of workstations Regularity Computing libraries Communication libraries Hpf compilers Data-parallelism