Grid5000:Home

Grid'5000 is a precursor infrastructure of SLICES-RI, Scientific Large Scale Infrastructure for Computing/Communication Experimental Studies. Content on this website is partly outdated. Technical information remains relevant.

Grid'5000 Grid'5000 is a large-scale and flexible testbed for experiment-driven research in all areas of computer science, with a focus on parallel and distributed computing, including Cloud, HPC, Big Data and AI. Key features: provides access to a large amount of resources: 15000 cores, 800 compute-nodes grouped in homogeneous clusters, and featuring various technologies: PMEM, GPU, SSD, NVMe, 10G and 25G Ethernet, Infiniband, Omni-Path highly reconfigurable and controllable: researchers can experiment with a fully customized software stack thanks to bare-metal deployment features, and can isolate their experiment at the networking layer advanced monitoring and measurement features for traces collection of networking and power consumption, providing a deep understanding of experiments designed to support Open Science and reproducible research, with full traceability of infrastructure and software changes on the testbed a vibrant community of 500+ users supported by a solid technical team Read more about our teams, our publications, and the usage policy of the testbed. Then get an account, and learn how to use the testbed with our Getting Started tutorial and the rest of our Users portal. Published documents and presentations: Presentation of Grid'5000 (April 2019) Report from the Grid'5000 Science Advisory Board (2018) Older documents: Slides from Frederic Desprez's keynote at IEEE CLUSTER 2014 Report from the Grid'5000 Science Advisory Board (2014) Grid'5000 is supported by a scientific interest group (GIS) hosted by Inria and including CNRS, RENATER and several Universities as well as other organizations. Inria has been supporting Grid'5000 through ADT ALADDIN-G5K (2007-2013), ADT LAPLACE (2014-2016), and IPL HEMERA (2010-2014).

Current status (at 2026-09-21 03:44): 6 current events, None planned (details)

Random pick of publications

Five random publications that benefited from Grid'5000 (at least 3007 overall):

• Georges da Costa. Hardware and application aware performance, power and energy models for modern HPC servers with DVFS. Sustainable Computing : Informatics and Systems, 2025, 46, pp.101106. 10.1016/j.suscom.2025.101106. hal-04983485 view on HAL pdf
• Neuville-Burguière Benjamin, Huet Fabrice, Baude Françoise. Latency Aware and ressource-efficient bin pack autoscaling for distributed event queues : An application to microservices architectures. Conférence francophone d'informatique en Parallélisme, Architecture et Système (COMPAS 2026), Jun 2026, Anglet, France. hal-05659911 view on HAL pdf
• Vladimir Ostapenco, Laurent Lefèvre, Anne-Cécile Orgerie, Benjamin Fichel. Exploring RAPL as a Power Capping Leverage for Power-Constrained Infrastructures. ICA3PP 2024 - 24th International Conference on Algorithms and Architectures for Parallel Processing, Oct 2024, Macau SAR, China. pp.1-10. hal-04742418 view on HAL pdf
• Florian Rascoussier, Romain Billot, Lina Fahed, Christine Solnon. Impact of Scaling and Rounding on Metaheuristic Performance for the Vehicle Routing Problem with Time Windows. 16th Metaheuristics International Conference, 2026, Ischia, Italy. hal-05646952 view on HAL pdf
• Anas Mokhtari, Baptiste Jonglez, Thomas Ledoux. Towards Digital Sustainability: Involving Cloud Users as Key Players. IC2E 2024 - 12th IEEE International Conference on Cloud Engineering, Sep 2024, Paphos, Cyprus. pp.126-132, 10.1109/IC2E61754.2024.00021. hal-04633237 view on HAL pdf

Latest news

Debian 13 "Trixie" std environment is now the default environment on (almost) all nodes

Dear users,

We are pleased to announce that Debian 13 (Trixie) std environment is

now the default environment on most nodes. See `kaenv3 -l debian%13%` for a list of available variants.

There is some significant evolutions, particularly regarding the

environment modules. Indeed, some components that were included in the environment images are now provided by modules loading, including CUDA, ROCm and OpenMPI.

Node network interface now have both IPv4 and (public) IPv6 addresses

by default, which should have no impact on most uses.

For more informations, see the Release Notes, available at

https://www.grid5000.fr/w/Debian_13_Trixie_environments_release_notes

As previous default environment move, frontend of every sites have been reinstalled in Debian 13 too.

If you have any questions or problems, please contact <support-staff@lists.grid5000.fr>

Best regards,

Nicolas Perrin, for Abaca and Grid’5000/SLICES-FR

Inria, DSI Service Production

-- Grid'5000 Team 14:54, 26 Aug 2026 (CEST)

Some clusters under Debian 13 "Trixie" environment by default and migration maintenance

Dear users,

The default standard environment has changed to Debian 13 for some

selected clusters over the past few weeks.

Here they are:

grenoble : vercors4 (3 nodes)

lyon  : hydra (4 nodes)

sophia  : esterel10 (3 nodes)

nancy  : grouille (2 nodes)

louvain  : spirou (8 nodes)

This is a key step towards the widespread deployment of Debian13. Most

clusters will be switched to a Debian 13 default environment Wednesday 26 August 2026.

If you need to, all the variants images (min, nfs, big) are available

for deployment, see `kaenv3 -l debian%13%` on frontends to list them.

Remember that there is a significant change concerning the modules. See

the preview announce : https://www.grid5000.fr/w/News#Upcoming_changes_to_modules

Best regards,

For Abaca and Grid'5000/SLICES-FR,

Nicolas Perrin

-- Grid'5000 Team 16:00, 02 Jul 2026 (CEST)

Upcoming changes to modules

Hello everyone,

Let's start with a quick TLDR, details on the rationale and implementation are available below: new modules will be available with the new standard environment, and are already live for testing (but not activated by default).

If you want to test them you need to run the following commands:

unset MODULEPATH

module use /grid5000/guix-modules/x86_64/latest
/grid5000/spack/module-others

Now for more details: for the past months we have been working on updating the standard environment and the way we provide modules.

The current way uses Spack under the hood, and is tightly bound to the underlying operating system.

It's been proven to be quite a burden for the team, and therefore we are changing the way we manage modules to:

- use a solution oblivious to the Linux flavor; - have means to update software versions automatically, and a clear release cycle; - actually have something reproducible; - be able to automatically test our most sensitive modules when they change (on both OAR and SLURM clusters).

Under the hood we switched to Guix to manage them²; it will be totally transparent for you.

The upcoming modules are located in `/grid5000/guix-modules/x86_64`, and you can try them today!

In order to use them you need to perform the following commands¹:

unset MODULEPATH

module use /grid5000/guix-modules/x86_64/latest

The list of modules for the `latest` release is available here:

https://api.grid5000.fr/explorer/software.

We know it's missing a few software compared to the current modules, we've tracked them here.

These should b...

Changes to VS Code and AI Extensions Usage on Frontend

Recently, we have observed a critical increase in resource consumption (CPU and memory) on these nodes. This is primarily caused by VS Code Server (or similar) instances and associated AI-assisted coding extensions (such as Copilot, Tabnine, or local LLM agents) running directly on the frontend.

As a reminder, frontends are strictly dedicated to lightweight tasks: code editing, file management, and job submission. Running heavy background processes or AI agents on these shared machines degrades performance for the entire community and risks crashing the machines. Frontend are not sized for heavy code/system compilation/build either. Heavy tasks must be run on reserved nodes.

What is changing:

- ban on frontend: running VS Code Server (or similar), AI extensions, or any background development agents directly on the frontend will shortly be prohibited. - automated cleanup: we will actively monitor these nodes. Any unauthorized, resource-intensive processes or persistent VS Code servers found running on the login nodes will be terminated without prior warning.

How to continue using VS Code and AI tools?:

We fully understand that these tools could be essential for your work. Therefore, this usage is completely permitted and supported on the compute nodes.

To use VS Code and your AI agents properly, you must schedule an interactive session via the batch scheduler (OAR). You can do this by:

- Requesting an interactive allocation using oarsub -I.

- Tunneling your VS Code Remote-SSH connection directly to the allocated compute node instead of the frontend

This ensures you have dedicated resources for your AI tools without impacting other users.

-- Grid'5000 Team 16:00, 15 Jun 2026 (CEST)

---

Read more news

Grid'5000 sites

Grenoble Lille Luxembourg Louvain

Lyon Nancy Nantes Rennes

Sophia-Antipolis Strasbourg Toulouse

Current funding

INRIA	CNRS
Universities IMT Atlantique Université Grenoble Alpes, Grenoble INP Université Rennes 1, Rennes Institut National Polytechnique de Toulouse / INSA / FERIA / Université Paul Sabatier, Toulouse Université Bordeaux 1, Bordeaux Université Lille 1, Lille École Normale Supérieure, Lyon	Regional councils Aquitaine Auvergne-Rhône-Alpes Bretagne Champagne-Ardenne Provence Alpes Côte d'Azur Hauts de France Lorraine