Stephen Trofinoff (CSCS) presents how to schedule GRES resources with SLURM.
Blog
Posts Tagged ‘Forum’
hpc-ch Forum on GPU – Video on Scheduling GRES resources With SLURM
Wednesday, November 23rd, 2011Agenda of hpc-ch Forum on Scheduler and Queuing Systems
Thursday, May 12th, 2011More than 30 members of the hpc-ch will participate to the next forum on «Scheduler and Queuing Systems» next Wednesday 18 May 2011. The forum will be hosted by PSI Paul Scherrer Institut. We are looking forward meeting you all and visit the Swiss Light Source (SLS) at the Paul Scherrer Institut.
Agenda
9:45 -10:00 Coffee and registration
10:00 -10:30 Greeting and introduction
- Welcome and Introduction; Derek Feichtinger (PSI) and Michele De Lorenzi (CSCS)
- Overview hpc-ch activities; Michele De Lorenzi (CSCS)
10:30 – 11:15 Internal Keynote presentation
- SGE+ARCO; Fabio Martinelli (PSI)
11:15 – 12:15 Members short presentations
- Using SLURM on XT, XE, and future Cray systems – We describe porting the open-source SLURM resource manager to the Cray interface; and report on experiences of using it on our main production platform, as well as several development systems of a heterogeneous multi-cluster environment. Since some of these systems are GPU-based, we also discuss issues in extending the existing interface to future systems, in order to take advantage of SLURMs cutting-edge GPU support; Kris Howard (CSCS)
- Moab for Grid Computing; Miguel Gila (CSCS)
12:15 – 13:00 Lunch in the foyer in front of the Auditorium
13:00 -14:30 Guided tour through SLS
14:30 – 15:30 Members short presentations
- An introduction to OGS/OGE/SGE (Open Gridschedular, Oracle Grid Engine, Sun Gridengine); Christian Bolliger (University of Zurich)
- SuperB: Pushing the limits of Torque and Maui - How to federate clusters while keeping them independent; Vittoria Rezzonico and Matteo Guglielmi (EPF Lausanne)
15:45 -17:00 Members short presentations
- Shedding some light into the black box – The batch system of Brutus is designed to work like a black box. This presentation will shed some light into the box, explain the basic ideas behind it and how they are implemented in Platform LSF; Olivier Byrde (ETH Zurich)
- The Meta-Scheduler ARC and gLite in the context of the Swiss Grid activities; Res Aeschlimann (SWITCH)
17:00 Farewell and end of the meeting
New Date for Forum on Scheduler and Queuing Systems
Thursday, January 27th, 2011The hpc-ch Forum on Scheduler and Queuing Systems at PSI had to be moved from May 19th, 2011 to
Wednesday, May 18th 2011, 10:00-17:30
because of organizational constraints. We hope, you will still be able to attend the event.
hpc-ch Events in 2011
Wednesday, December 15th, 2010We are happy to announce the events of the hpc-ch community for the year 2011. Like in 2010 we will organize two Forums where we will discuss important topics for our community. Both topics have been selected by the community representatives during the last Forum on parallel file systems for HPC hosted by ETH Zurich.
The first Forum on Scheduler and Queuing Systems will be hosted by PSI Paul Scherrer Institut in Villigen (next to Baden). Our colleagues from PSI (thanks Derek!) promised to organize a visit to their research infrastructure (possibly the Swiss Light Source SLS).
The second Forum will be dedicated to the Use of GPU and Accelerators for HPC and will be hosted by the University of Basel. Again we will have to possibility to have look to the research activities of the university.
- Thursday 19 May 2011, 10:00 - 17:30
Forum on Scheduler and Queuing Systems, hosted by PSI – Paul Scherrer Institut - Thursday 27 October 2011, 10:00 - 17:00
Forum on the Use of GPU and Accelerators for HPC, hosted by Universität Basel
The community representatives will receive in due time by e-mail a Call for Presentations and Participation to each Forum.
We would also like to remember you an additional possibility to meet during the HPC Advisory Council Switzerland (registrations are already open).
- Monday 21 March 2011 9:00 - Tuesday 23 March 2010 - 17:00
HPC Advisory Council Switzerland Workshop 2011, hosted by CSCS – Swiss National Supercomputing Centre
Videos of Forum on Parallel File Systems for HPC now Online
Wednesday, November 10th, 2010The recordings of the last hpc-ch community forum on Parallel file systems for HPC are now online. More than 40 colleagues from all over Switzerland attended the forum. The videos of the 9 presentations give an overview of the state of the art in parallel file systems in Switzerland: The technologies being used, their pros and cons and the future plans.
You can go directly to the web page with the videos »
or continue reading this post to get a list of all presentations (click on the picture with the selected speaker to start the video).
Overview of HPC at ETH Zurich, Olivier Byrde (ETH Zurich)
Since ETH Zurich hosted the forum, we had an introduction to the HPC systems being deployed at ETH Zurich followed by a site visit. If you want an overview of Brutus, the large cluster running at ETH Zurich, than have a look to this presentation. If you are more interested on parallel file systems, than you can skip to the next one.
Parallel File System for Brutus; Teodoro Brasacchio and Adrian Ulrich (ETH Zürich/CSCS)
GPFS Use at the PSI Swiss Light Source; Heiner Billichs (PSI)
Parallel Files Systems at CSCS; Davide Tacchella (CSCS)
Disk Pool Manager Grid Storage Element; Szymon Gadomski (University of Geneva)
The software distributes abstract directories over multiple physical file servers and provides similar functionality to a distributed file system. We will present our experiences of the last 1.5 years and our performance measurements.
Finding the perfect filesystem for life sciences an impossible mission?; Roberto Fabbretti and Volker Flegel (Vital IT)
Experiences with Lustre in Bern; Gianfranco Sciacca (University of Bern)
Lustre based I/O on Cray XT systems; Jason Temple (CSCS)
The planned European OFS Cooperative Society SCE; Christian Bolliger (University of Zurich)
After Oracle bought Sun and thus the rights on Lustre, the future development of Lustre on Linux for HPC is not ensured. Therefore several initiatives have been started to bundle the community efforts on Lustre. Among these is the plan to set up the “European-OFS-Cooperative-SCE” before end of the year. Some of the bigger HPC centers, like the Jülich Supercompting Center and the Barcelona Supercomputing Center as well as HPC vendors like HP, Cray and Bull and service companies are showing a deep interest in this initiative.
