News

Monday, 03 March 2014

22,260,000 core-hours on PRACE Tier-0 clusters allocated to CÉCI users. Be the next one!

Jean-Christophe Charlier and Nicolas Leconte (UCL), and Samir Sid and Vincent Terrapon (ULg) have been allocated time on Tier-0 clusters, respectively in the context of the 8th Regular Call and for a preparatory access. Congratulation to them.

As the 9th Call for Proposal is now open, do not hesitate to apply.

Thursday, 27 February 2014

NIC4 installed at ULg

The sixth CECI cluster, NIC4 has entered production. It is installed at ULg. NIC4 features 120 compute nodes with two 8-cores Intel E5-2650 processors at 2.0 GHz and 64 GB of RAM (4 GB/core), interconnected with a QDR Infiniband network (2:1 blocking factor), and having exclusive access to a fast 144 TB FHGFS parallel filesystem. From today, you can access it with your CÉCI credentials.

Thursday, 21 November 2013

[French] Appel à communication: ComPAS'2014, Neuchatel, Suisse, 22-25 avril 2014

ComPAS est la Conférence d'informatique en Parallélisme, Architecture et Système, francophone, et orientée Jeunes Chercheurs. Depuis 2013, elle regroupe en une identité unique les trois précédentes conférences francophones RenPar (Rencontres francophones du Parallélisme), SympA (Symposium en Architectures nouvelles de machines) et CFSE (Conférence Française en Systèmes d'Exploitation). ComPAS'2014 aura lieu du 22 au 25 avril 2014 à Neuchâtel, en Suisse Romande.

See more information here.

Thursday, 03 October 2013

200.000 core-hours on PRACE Tier-0 clusters allocated to a CÉCI user

Yvan Notay and his team from ULB have been granted 100.000 core-hours on Juqueen, and 50.000 more on Hermit and Curie each, to test the scalability of their code AGMG, resulting from an application to a PRACE Preparatory Access call Type A. Congratulation to them!

Tuesday, 17 September 2013

Training sessions to start mid-October

The CISM and the CÉCI jointly offer, this year again, a complete training for super-computer users from a basic introduction to Linux to complex parallel computations with MPI. Registration is free but mandatory. More info on the training page.

Wednesday, 11 September 2013

VEGA installed at ULB

The HPC cluster 'Vega' is now installed at the Université libre de Bruxelles. Vega is primarily dedicated to many-cores jobs (each node has 64 cores) and to batches of many single-core jobs. It comprises 44 nodes, each with four AMD Opteron 6272, 256 GB of RAM and 250GB of local harddisk (yes, more RAM than disk space). Vega also has a fast 70TB GPFS global scratch space.

Wednesday, 04 September 2013

8th PRACE Regular Call for Proposals open

The 8th PRACE Regular Call for Proposals will be open from September 3rd, 2013 until October 15th, 2013, noon, 12:00 Brussels Time.

More info at http://www.prace-ri.eu/PRACE-Project-Access

Sunday, 18 August 2013

Slurm update on Hercules, Lemaitre2 and Hmem

During the maintenance operations on Hercules, Lemaitre2 and Hmem earlier this Summer, Slurm was updated to version 2.6.0.

Interesting changes include:

  • Support for job arrays
  • Added support for job profiling (CPU, memory, power consumption, Lustre use and Infiniband use)
  • Much faster throughput for job step execution with --exclusive option
  • Advanced reservations with hostname and core counts now supports asymetric reservations

See more information here.

Sunday, 07 July 2013

HERCULES installed at UNamur

The HPC cluster 'Hercules' is available for use. Hercules is located at the University of Namur and operated by the 'Plateforme Technologique de Calcul Intensif' (PTCI). Hercules is primarily designed for users who need to run shared memory parallel jobs (OpenMP or Pthreads) or multiple resource-intensive sequential jobs. It comprises 32 Intel Sandy Bridge nodes, each with two 8-core processors and 64 or 128 GB of RAM. One partition is available with no time limit.

Sunday, 12 May 2013

PRACE Call for Proposal

PRACE has issued a Call for Proposals for DECI-11 (Tier-1): ( deadline: 10th June 2013, 12:00 CEST ; stake: 1 year allocation starting: 1st Nov, 2013 ; resources: all Tier-1 machines)

Let us know if you respond and participate!

Press Coverage

Tuesday, 21 January 2014

Media coverage of the Tier1 inauguration

The CÉCI has often been referred to in the press coverage of the inauguration of the Tier-1 Supercalculateur at Cenaero on Monday, January the 20th.

The following news bulletins excerpts feature members of the CÉCI being interviewed. They all date Jan 20th.

Some articles:

Thursday, 10 October 2013

CanalC news topic about UNamur and Hercules

Canal C's news bulletin from October 9 features UNamur's cluster Hercules.

See the video here.

Friday, 20 September 2013

La Libre article mentioning Hercules

La Libre has published an article entitled 'Les Belges bousculent l'icône Einstein' where the Hercules cluster is mentioned.

See the article here.

Sunday, 01 September 2013

InsideHPC article about Vega

InsideHPC and Scientific Computing World have released an article about Vega, the new CÉCI cluster installed at ULB. The article also mentions other CÉCI clusters.

See the article here or here.

Saturday, 22 June 2013

La Libre article mentioning the CÉCI

La Libre has published an article about supercomputers, featuring UNamur CÉCI members.

See the article here.

Monday, 22 April 2013

La Libre article about UNamur's cluster inauguration.

La Libre has a short article about the latest CÉCI scientific day, held in Namur on April 19th, 2013, where the new CÉCI cluster was inaugurated.

See the article here.

Tuesday, 01 May 2012

Article about CÉCI and CISM in 'La Quinzaine'

La Quinzaine, the internal magazine of UCLouvain, publishes an article about the CISM and the CÉCI in its May issue.

See the article here.

Friday, 10 February 2012

Second Datanews article about CÉCI this year

The latest issue of Datanews offers a one-page article about the CÉCI. Based on an interview with Thomas Keutgen, manager at the supercomputing centre of UCL, the article details the CÉCI initiative, organization and roadmap. The article highlights the fact that, by 2013, the CÉCI clusters should represent a total of 6000 to 7000 compute cores.

See the article here.

Friday, 27 January 2012

Datanews article about CÉCI

Datanews publishes an article about Lemaitre2, the new CÉCI cluster installed at UCL. The article, titled "L'UCL installe 14 Tflops", describes the new cluster and compares its features to the olders clusters from UCL. The article also brievely presents the CÉCI and its organisation.

See the article here.

Monday, 20 December 2010

Roundup for the year 2010

  • 22 November 2010 : Article in "La Libre Belgique" about the (lack of) funding of the F.R.S.-FNRS, citing the CÉCI as an example.

  • 17 November 2010 : Two mentions of the CÉCI agreement signature on the FNRS news. Read the first one and the second one.

  • 16 November 2010 : TV report by Canal Zoom about the signature of the CÉCI consoriumt agreement.

  • 12 November 2010 : Article in "Le Vif" about the creation of the CÉCI consortium.

© 2013 CÉCI.