Grid5000:Home: Difference between revisions

From Grid5000
Jump to navigation Jump to search
No edit summary
No edit summary
 
(70 intermediate revisions by 10 users not shown)
Line 3: Line 3:
|- valign="top"
|- valign="top"
|bgcolor="#f5fff5" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
|bgcolor="#f5fff5" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
[[Image:Logo.png|left]]
[[Image:g5k-backbone.png|thumbnail|260px|right|Grid'5000]]
<br>
'''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 and Big Data and AI.'''
''Grid'5000 is a scientific instrument supporting experiment-driven research in all areas of computer science, including high performance computing, distributed computing, networking and big data.'' <br>
[[media:seminaire_intro.pdf|Download the latest general introduction]], or a [https://www.grid5000.fr/screencast/index.html screencast of recent webUI developments]
|}
{{#status:0|0|0|http://bugzilla.grid5000.fr/status/upcoming.json}}
== Latest updates from Grid'5000 users ==
* '''Publications'''
{{#publications:3||**}}
==Latest news==
=== Journée Aramis ===


Le 16/04 aura lieu la journée Aramis "Usages et services du cloud dans notre environnement enseignement supérieur et recherche" avec une présentation de Grid'5000 par Simon Delamare.
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


Cette journée est webcastée par la cellule Webcast du CC-IN2P3. Vous pouvez donc suivre en direct les [http://webcast.in2p3.fr/live/pleiniere_aramis_2014 présentations]. Le programme de la journée est disponible [http://aramis.resinfo.org/wiki/doku.php?id=pleniaires:pleniere17avril2014 ici]
<br>
Read more about our [[Team|teams]], our [[Publications|publications]], and the [[Grid5000:UsagePolicy|usage policy]] of the testbed. Then [[Grid5000:Get_an_account|get an account]], and learn how to use the testbed with our [[Getting_Started|Getting Started tutorial]] and the rest of our [[:Category:Portal:User|Users portal]].


=== [[Grid5000:UserCharter|Grid'5000 Charter]] revised ===
<b>Grid'5000 is merging with [https://fit-equipex.fr FIT] to build the [http://www.silecs.net/ SILECS Infrastructure for Large-scale Experimental Computer Science]. Read [http://www.silecs.net/wp-content/uploads/2018/04/Desprez-SILECS.pdf an Introduction to SILECS] (April 2018)</b>
The Grid'5000 charter has been reworded and simplified so as to be more inclusive about possible uses. The rules to ensure availability during the day for any user and during the night for large resource usage have not been changed.
=== [[Grid5000:school2014|Grid'5000 school 2014]] announced ===
Organized in Lyon, from June 16th to June 19th 2014, this [[Grid5000:school2014|spring school]] will bring together, but is not limited to, Grid'5000's users, technical team and executive committee for 4 days of tutorials and talks focusing on best-practices and results. Presentations and practical sessions will cover both basic usage of the platform, for new users, or potential users of Grid'5000 and advanced and new usage of the platform, for current users. Deadline to submit presentation proposals is March 29th, 2014.


=== Grid'5000 users win second prize at CCGRID 2013's SCALE challenge ===
<br>
Snooze based entry running on Grid'5000 entry wins [http://www.pds.ewi.tudelft.nl/ccgrid2013/awards/ 2nd prize] at [http://www.pds.ewi.tudelft.nl/ccgrid2013/ CCGrid 2013] [http://www.pds.ewi.tudelft.nl/ccgrid2013/calls/scale-challenge/ SCALE challenge]: well done Matthieu and Anne-Cécile for defending the entry titled ''Scalability of the Snooze Autonomic Cloud Management System'' by Eugen Feller, Christine Morin, Matthieu Simonin, Anne-Cécile Orgerie, and Yvon Jégou.
Recently published documents and presentations:
* [[Media:Grid5000.pdf|Presentation of Grid'5000]] (April 2019)
* [https://www.grid5000.fr/mediawiki/images/Grid5000_science-advisory-board_report_2018.pdf Report from the Grid'5000 Science Advisory Board (2018)]


=== Grid'5000 users finalists of the SCALE'2013 challenge ===
Older documents:
Two submissions (out of five) from Grid'5000 users took part in the final of the international SCALE'2013 challenge (held with CCGrid'2013):
* [https://www.grid5000.fr/slides/2014-09-24-Cluster2014-KeynoteFD-v2.pdf Slides from Frederic Desprez's keynote at IEEE CLUSTER 2014]
* D. Balouek, A. Lèbre, F. Quesnel Flauncher and DVMS -- Deploying and Scheduling Thousands of Virtual Machines on Hundreds of Nodes Distributed Geographically
* [https://www.grid5000.fr/ScientificCommittee/SAB%20report%20final%20short.pdf Report from the Grid'5000 Science Advisory Board (2014)]
* Eugen Feller, Christine Morin, Matthieu Simonin, Anne-Cécile Orgerie, and Yvon Jégou -- Scalability of the Snooze Autonomic Cloud Management System.


The first proposition presents the deployment and scheduling of thousands of virtual machines, conducted with the Flauncher and DVMS frameworks, across the Grid'5000 testbed.
<br>
The frameworks have been able to deploy and schedule up to 10000 VMs during the tests. This research has been conducted in the context of the INRIA Hemera initiative.
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|HEMERA]] (2010-2014).
 
|}
The second proposition presents the Snooze architecture and focus on the following aspects :
* System set up scalability and resources consumption,
* Self-healing capabilities.
The deployment of Snooze were distributed over several sites of the Grid'5000 testbed.
The framework has been able to start 11000 of system services, recover thousands of failures
and used to launch large hadoop/mapreduce experiments.
This research has been conducted in the context of the INRIA Snooze ADT.
=== Grid'5000 tutorial during ComPAS'2013 ===
The [http://compas2013.inrialpes.fr/ ComPAS'2013 conference] (replacing RenPar, SympA and CFSE), to be held in Grenoble between January 15th and 18th, will feature a Grid'5000 tutorial.
----
[[Grid5000:News|read more news]]


<br>
<br>
==Grid'5000 at a glance==
{{#status:0|0|0|http://bugzilla.grid5000.fr/status/upcoming.json}}
<!-- [[Image:site_map.png|thumbnail|128px|right|Grid'5000 sites]] -->
<br>
[[Image:renater5-g5k.jpg|thumbnail|200px|right|Grid'5000]]


== Random pick of publications ==
{{#publications:}}


* '''Grid'5000''' is a scientific instrument for the study of large scale parallel and distributed systems. It aims at providing a '''highly reconfigurable, controlable and monitorable experimental platform''' to its users. The initial aim (circa 2003) was to reach 5000 processors in the platform. It has been reframed at 5000 cores, and was reached during winter 2008-2009.
==Latest news==
* The infrastructure of Grid'5000 is geographically distributed on different sites hosting the instrument, initially 9 sites in France (10 since 2011). Porto Alegre, Brazil is now officially becoming the first site abroad.
<rss max=4 item-max-length="2000">https://www.grid5000.fr/rss/G5KNews.php</rss>
----
[[News|Read more news]]


===Sites:===
=== Grid'5000 sites===
{|width="75%" cellspacing="3"  
{|width="100%" cellspacing="3"  
|- valign="top"
|- valign="top"
|width="33%" bgcolor="#f5f5f5" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
|width="33%" bgcolor="#f5f5f5" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
* [[Bordeaux:Home|Bordeaux]]
* [[Grenoble:Home|Grenoble]]
* [[Grenoble:Home|Grenoble]]
* [[Lille:Home|Lille]]
* [[Lille:Home|Lille]]
Line 67: Line 53:
* [[Lyon:Home|Lyon]]
* [[Lyon:Home|Lyon]]
* [[Nancy:Home|Nancy]]
* [[Nancy:Home|Nancy]]
* [[Nantes:Home|Nantes]] (soon)
* [[Nantes:Home|Nantes]]
* [[Reims:Home|Reims]]
|width="33%" bgcolor="#f5f5f5" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
|width="33%" bgcolor="#f5f5f5" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
* [[Rennes:Home|Rennes]]
* [[Rennes:Home|Rennes]]
Line 75: Line 60:
|-
|-
|}
|}
[[Image:Software layers.png|thumbnail|271px|left|Grid'5000 allows experiments in all these software layers]]
* '''Grid'5000''' is a research effort developing a '''large scale nation wide infrastructure for large scale parallel and distributed computing research'''.
* '''19 [[Grid5000:Laboratories|laboratories]]''' are involved in France with the objective of providing the community a testbed allowing experiments in all the software layers between the network protocols up to the applications.
The current plans are to extend from the 9 initial sites each with 100 to a thousand PCs, connected by the [http://www.renater.fr RENATER] Education and Research Network to a bigger platform including a few sites outside France not necessarily connected through a dedicated network connection. Sites in Brazil and Luxembourg should join shortly, and Reims has now joined.
All sites in France are connected to [http://www.renater.fr RENATER] with a 10Gb/s link, except Reims, for the time linked through a 1Gb/s
This high collaborative research effort is funded by INRIA, CNRS, the Universities of all sites and some regional councils.
== ALADDIN-G5K : ensuring the development of '''Grid'5000''' ==
For the 2008-2012 period, Engineers ensuring the development and day to day support of the infrastructure are mostly provided by INRIA, under the ''ADT ALADDIN-G5K''  initiative.
==[[Hemera|HEMERA: Demonstrating ambitious up-scaling techniques on '''Grid'5000''']] ==
[[Hemera|Héméra]] is an INRIA Large Wingspan project, started in 2010, that aims at demonstrating ambitious up-scaling techniques for large scale distributed computing by carrying out several dimensioning experiments on the Grid’5000 infrastructure, at animating the scientific community around Grid’5000 and at enlarging the Grid’5000 community by helping newcomers to make use of Grid’5000.
== Initial Rationale==
'''The foundations of Grid'5000''' have emerged from a thorough analysis and numerous discussions about methodologies used for scientific research in the Grid domain. A report presents the [http://www-sop.inria.fr/aci/grid/public/Library/rapport-grid5000-V3.pdf rationale for Grid'5000].
In addition to theory, simulators and emulators, there is a strong need for '''large scale testbeds''' where real life experimental conditions hold. '''The size of Grid'5000''', in terms of number of sites and number of processors per site, was established according to the scale of the experiments and the number of researchers involved in the project.


== Current funding ==
== Current funding ==
As from June 2008, INRIA is the main contributor to [[Grid5000:Funding|Grid'5000 funding]].  
As from June 2008, Inria is the main contributor to [[Grid5000:Funding|Grid'5000 funding]].  
{|width="100%" cellspacing="3"
{|width="100%" cellspacing="3"
|-
|-
Line 114: Line 70:
| width="50%" bgcolor="#f5f5f5" valign="top" align="center" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
| width="50%" bgcolor="#f5f5f5" valign="top" align="center" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
===CNRS===
===CNRS===
[[Image:CNRS-filaire-MonoBleu.gif|100px]]
[[Image:CNRS-filaire-Quadri.png|125px]]
|-
|-
| width="50%" bgcolor="#f5f5f5" valign="top" align="center" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
| width="50%" bgcolor="#f5f5f5" valign="top" align="center" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
===Universities===
===Universities===
University Joseph Fourier, Grenoble<br/>
IMT Atlantique<br/>
University of Rennes 1, Rennes<br/>
Université Grenoble Alpes, Grenoble INP<br/>
Université Rennes 1, Rennes<br/>
Institut National Polytechnique de Toulouse / INSA / FERIA / Université Paul Sabatier, Toulouse<br/>
Institut National Polytechnique de Toulouse / INSA / FERIA / Université Paul Sabatier, Toulouse<br/>
University Bordeaux 1, Bordeaux<br/>
Université Bordeaux 1, Bordeaux<br/>
University Lille 1, Lille<br/>
Université Lille 1, Lille<br/>
Ecole Normale Supérieure, Lyon<br/>
École Normale Supérieure, Lyon<br/>
Université de Reims Champagne-Ardenne, Reims<br/>
| width="50%" bgcolor="#f5f5f5" valign="top" align="center" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
| width="50%" bgcolor="#f5f5f5" valign="top" align="center" style="border:1px solid #cccccc;padding:1em;padding-top:0.5em;"|
===Regional councils===
===Regional councils===
Aquitaine<br/>
Aquitaine<br/>
Auvergne-Rhône-Alpes<br/>
Bretagne<br/>
Bretagne<br/>
Champagne-Ardenne<br/>
Champagne-Ardenne<br/>
Provence Alpes Côte d'Azur<br/>
Provence Alpes Côte d'Azur<br/>
Nord Pas de Calais<br/>
Hauts de France<br/>
Lorraine<br/>
Lorraine<br/>
|}
|}

Latest revision as of 10:29, 26 October 2023

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 and 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.

Grid'5000 is merging with FIT to build the SILECS Infrastructure for Large-scale Experimental Computer Science. Read an Introduction to SILECS (April 2018)


Recently published documents and presentations:

Older documents:


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 2025-05-05 17:14): 1 current events, 9 planned (details)


Random pick of publications

Five random publications that benefited from Grid'5000 (at least 2758 overall):

  • Hee-Soo Choi, Priyansh Trivedi, Mathieu Constant, Karën Fort, Bruno Guillaume. Au-delà de la performance des modèles : la prédiction de liens peut-elle enrichir des graphes lexico-sémantiques du français ?. Actes de JEP-TALN-RECITAL 2024. 31ème Conférence sur le Traitement Automatique des Langues Naturelles, volume 1 : articles longs et prises de position, Jul 2024, Toulouse, France. pp.36-49. hal-04623008 view on HAL pdf
  • Alaaeddine Chaoub. Deep learning representations for prognostics and health management. Computer Science cs. Université de Lorraine, 2024. English. NNT : 2024LORR0057. tel-04687618 view on HAL pdf
  • Fatmir Asllanaj, Sylvain Contassot-Vivier, Guilherme C Fraga, Francis H.R. França, Roberta J.C. da Fonseca. New gas radiation model of high accuracy based on the principle of weighted sum of gray gases. Journal of Quantitative Spectroscopy and Radiative Transfer, 2024, 315, pp.108887. 10.1016/j.jqsrt.2023.108887. hal-04375649 view on HAL pdf
  • Maxime Méloux, Christophe Cerisara. Novel-WD: Exploring acquisition of Novel World Knowledge in LLMs Using Prefix-Tuning. 2023. hal-04269919 view on HAL pdf
  • Imran Ahamad Sheikh, Emmanuel Vincent, Irina Illina. Training RNN Language Models on Uncertain ASR Hypotheses in Limited Data Scenarios. Computer Speech and Language, 2024, 83, pp.101555. 10.1016/j.csl.2023.101555. hal-03327306v2 view on HAL pdf


Latest news

Rss.svgCluster chirop is now in the default queue of Lille with energy monitoring.

Dear users,

We are pleased to announce that the Chirop[1] cluster of Lille is now available in the default queue.

This cluster consists of 5 HPE DL360 Gen10+ nodes with:

  • 2 CPU Intel Xeon Platinum 8358 (32 cores per CPU)
  • 512 GiB memory
  • 1*1.92TB SSD NVME + 2*3.84TB SSD
  • 2*25 Gbps Ethernet interface

    • Energy monitoring[2] is also available for this cluster[3], provided by newly installed Wattmetres (similar to those already available at Lyon).

    This cluster was funded by CPER CornelIA.

    [1] https://www.grid5000.fr/w/Lille:Hardware#chirop

    [2] https://www.grid5000.fr/w/Energy_consumption_monitoring_tutorial [3] https://www.grid5000.fr/w/Monitoring_Using_Kwollect#Metrics_available_in_Grid.275000

    -- Grid'5000 Team 16:25, 05 May 2025 (CEST)

    Rss.svgChange of default queue based on platform

    Until now, Abaca (production) users had to specify `-q production` when reserving Abaca resources with OAR.

    This is no longer necessary as your default queue is now automatically selected based on the platform your default group is associated to, as shown at https://api.grid5000.fr/explorer/selector/ and in the message displayed when connecting to a frontend.

    For SLICES-FR users, there is no change since the correct queue was already selected by default.

    Additionally, the "production" queue has been renamed to "abaca", although "production" will continue to work for the foreseeable future.

    Please note one case where this change may affect your workflow:

    When an Abaca user reserves a resource from SLICES-FR (a non-production resource), they must explicitly specify they want to use the SLICES-FR queue, which is called "default", by adding `-q default` the OAR command.

    -- Abaca Grid'5000 Team 10:10, 31 March 2025 (CEST)

    Rss.svgCluster "musa" with Nvidia H100 GPUs is available in production queue

    We are pleased to announce that a new cluster named "musa" is available in the production queue¹ of Abaca.

    This cluster has been funded by Inria DSI as a shared computing resource.

    It is accessible to all Abaca users. Users affiliated with Inria have access with the same level of priority, regardless of the research center to which they are attached.

    This cluster is composed of six HPE Proliant DL385 Gen11 nodes² with 2 AMD EPYC 9254 24-Core Processor, 512 GiB of RAM, 2 x Nvidia H100 NVL (94 GiB) with NVLink, one 6 TB SSD NVME and 25 Gbps Ethernet Connexion

    Please note that in order to share it efficiently, walltime is limited:

  • 6 hours for the first two nodes
  • 24 hours for the next two
  • 48 hours for the last two

    • The cluster "musa" is located at Sophia, hosted in the datacenter of Inria Centre at Université Côte d’Azur.

    ¹: https://api.grid5000.fr/explorer/hardware/sophia/#musa

    ²: the nodes are named musa-1, musa-2,.., musa-6

    -- Grid'5000 Team 13:30, 19 March 2025 (CEST)

    Rss.svgCluster "Hydra" is now in the testing queue in Lyon

    We are pleased to announce that the hydra[1] cluster of Lyon is now available in the testing queue.

    Hydra is a cluster composed of 4 NVIDIA Grace-Hopper servers[2].

    Each node features:

  • 1 Nvidia Grace ARM64 CPU with 72 cores (Neoverse-V2)
  • 1 Nvidia Hopper GPU
  • 512GB LPDDR5 memory
  • 96GB HBM memory
  • 1x1To SSD NVME + 1x1.92To SCSI disk

    • Due to its bleeding edge hardware, usual Grid'5000 environments are not supported by default for this cluster.

    (Hydra requires system environments featuring a Linux kernel >= 6.6). The default system on the hydra nodes is based on Debian11, but **does not provide functional GPU**. However, users may deploy the ubuntugh2404-arm64-big environment, which is similar to official Nvidia image provided for this machine and provides GPU support.

    To submit a job on this cluster, the following command may be used:

    oarsub -q testing -t exotic -p hydra

    This cluster is funded by INRIA and by Laboratoire de l'Informatique du Parallélisme with ENS Lyon support.

    [1] Hydra is the largest of the modern constellations according to Wikipedia: https://en.wikipedia.org/wiki/Hydra_(constellation)

    [2] https://developer.nvidia.com/blog/nvidia-grace-hopper-superchip-architecture-in-depth/

    -- Grid'5000 Team 16:10, 11 March 2025 (CEST)

    Read more news

    Grid'5000 sites

    Current funding

    As from June 2008, Inria is the main contributor to Grid'5000 funding.

    INRIA

    Logo INRIA.gif

    CNRS

    CNRS-filaire-Quadri.png

    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

    Retrieved from "https://www.grid5000.fr/mediawiki/index.php?title=Grid5000:Home&oldid=100740"