• Digital World and Robotics

"Mésocentre" of High Performance Computing of the University of Lille (Mésocentre de Lille)

Platform

Description

Our platform is dedicated to the validation of research projects in digital simulation, modeling, high performance computing, and artificial intelligence. We provide high performance computing facilities as well as our expertise in the field of scientific computing to process large volumes of data from experiments or simulations. We are open to all academic collaborations in Hauts-de-France and propose services in line with our expertise.

Contacts

  • Matthieu Marquillie
    Technical manager
  • Nouredine Melab
    Scientific manager

Informations

Bâtiment M4, Avenue Carl Gauss
Campus Cité scientifique, Université de Lille
59650 VILLENEUVE D'ASCQ

http://hpc.univ-lille.fr/

Chiffres clés

• Over 350 users per year
• Number of computing hours: more than 68 million hours per year
• Computing cluster: 340 Tflops composed of 6264 cores, 38 TB of RAM and 630 TB of storage.
• Computing cloud: 62 Tflops of power composed of 1436 cores, 17 TB of RAM and 215 TB of storage
• Storage: 2.8Po
• 21 Mesocenters partners in the equipex+ MesoNET
• 40 user laboratories
• Between 20 to 30 articles per year
• About 5 theses per year

Effectif

Effectif total : 3

Expertises

Skills

• Provision of supercomputers with user support
• Administration of supercomputers and storage systems
• Pooling of computing resources at different levels: local, regional, national and European
• Study of opportunities for and assistance in parallelization of HPC code
• Parallel programming training

Example(s) of projects

• ERC RheoMan 🡭
• ERC TimeMan 🡭

Discover the full list of publications here.

Collaborations/Partners/Scientific clients

• 32 laboratories of the University of Lille users + 8 regional laboratories
Study of opportunities for and assistance in parallelization of HPC code
• 21 partner mesocentres as part of the Equipex + MesoNET

Services offers

Applications sectors

  • IT / software
  • Science / Research

Services provided

• Three platforms (and their system & network administration) are available to users:
- Cluster computing: simulation, data processing, artificial intelligence, GPU
- Cloud computing: simulation, data processing, scientific portals hosting
- Storage plateform: scientific data storage
• Technical support (daily) for users and training in the use of the three platforms
• Technical expertise (code optimization, parallelization) and training in parallel programming (MPI and OpenMP)
• Support for teams in calls for projects (eg ERC P. Cordier), recommendation of hardware and software configuration, quotes, etc.

Training offers

• Training in the use of computing and/or storage resources (job submission on clusters, massive storage, etc.)
Technical support (daily) for users and training in the use of the three platforms
• Training in parallel programming: OpenMP and MPI

Consulting services

We provide consulting services related to all our areas of expertise:
• technical support to users
• study of opportunities for and help in the parallelization of high-performance computing code
• help in the use of national level calculation resources (IDRIS, CINES)

Equipments

We have three platforms:
• Computing cluster: 5,880 cores, 35 TB of RAM and 630 TB of storage for a total power of 332 TFlops
• Computing cloud: 2,444 physical cores, 19 TB of RAM and 215 TB of storage for a total power of 84 TFlops
• Long-term storage: 2 Po of storage

See the detailed list of equipment: http://hpc.univ-lille.fr/cluster-hpc-htc

Equipments

Nom Modèle Marque
Computing cluster Divers
Cloud computing Openstack/ Divers
Scientific storage Divers

Ecosystem

Affiliated institutions / organisations

Groups/Networks/Federations

Labellisations

Regional strategic areas of activity

  • Digital World and Robotics
    • Internet of things, cloud computing and big data