Team

Runtime

LaBRI, Inria Bordeaux - Sud-Ouest

High Performance Runtime Systems for Parallel Architectures

Presentation of the Team

The Inria RUNTIME team takes place in the high-performance and exa-scale parallel computing domain. It is interested in the study and design as well as the development and evaluation of the principles that will be at the heart of tomorrow's parallel runtime systems. More precisely, we intend to define, implement and validate a family of generic runtime systems that will provide a flexible and efficient foundation for building programming environments and applications in the intensive parallel computing field. These runtime systems will have to enable an efficient use of the large, heterogeneous and hierarchical clusters, based on NUMA and/or multicore nodes, even with accelerators such as GPUs.

A more detailed presentation of our research areas can be found in the 2012 activity report.

RUNTIME is headed by Professor Raymond Namyst. For a full list of members, see here.

News

  • Runtime is ending. It will be replaced with the Storm and TADaaM Inria teams.
  • 2015/09 - We're organizing the EuroMPI 2015 conference in Bordeaux.
  • 2014/03/27 - Bertrand Putigny defended his PhD thesis at ENSEIRB-MATMECA, University of Bordeaux.
  • 2014/01/06 - Christopher Haine is starting a PhD Thesis in the team
  • 2013/11/18 - The Runtime group was at SuperComputing (SC13) in Denver, presenting its HPC software at Inria booth (#2116).
  • 2013/11/14 - Sylvain Henry is defending his PhD thesis.
  • 2013/07/15 - Lilia Ziane Khodja is starting a post-doc position in the team.
  • 2013/06/24 - Runtime organized CEA-EDF-Inria HPC summer school on Programming Heterogeneous Parallel Architectures.
  • 2013/01/21 - Runtime gave a tutorial about affinities and locality with hwloc at the HIPEAC Conference in Berlin.
  • 2013/01/15 - Runtime gave tutorials on StarPU and hwloc at the French ComPAS conference.

Software Releases

  • 2014/12/15 - Stable release 1.5.4 of the Open-MX high-performance communication stack, available from here.
  • 2014/12/08 - Stable release candidate 1.10.1rc1 of Hardware Locality (hwloc), download it here!
  • 2014/02/24 - Stable release 1.1.1 of KNEM, the high-performance intra-node MPI communication module, download it here!
  • 2013/12/18 - Release 1.1.0 of StarPU, download it here!

Collaborations

Academic collaborations

Industrial Collaborations

Contacts

For any information, please contact Raymond NAMYST :