Forum on Concurrent Programming Models and Frameworks

Announcing Annual Concurrency Forum Meeting - April 1-2, 2014

Pere Mato Vila — Thu, 06 Mar 2014 10:12:24 +0000

The Annual Concurrency Forum Meeting will be held on April 1-2, 2014 at CERN. This meeting is the follow-up of the Annual Meeting that took place in February 2013 at Fermilab. We expect to have status reports from all the activities of the Concurrency Forum. The meeting will be followed by the HEP Software Collaboration Meeting in which the proposal for the creation of new software collaboration will be discussed.

TechLab project announcement

Pere Mato Vila — Tue, 23 Jul 2013 06:50:20 +0000

The IT department is preparing a new project together with PH/SFT aiming at improving the efficiency of our computing architecture and making better utilisation of the processors available today.

This best-effort activity is called Techlab and will gradually be made available over the coming months.
It consists of a several test systems, which users can book for a limited time, in order to improve their applications and evaluate the possible efficiency and performance gain, from both an optimised software environment and modern hardware.

A first set of advanced computing equipment (large multi-core node, Nvidia GPU) is now available for tests and possibly interested people should contact Romain Wartel <Romain.Wartel@cern.ch> for details.

More information on Techlab is available at https://twiki.cern.ch/twiki/bin/view/IT/TechLab

In order to receive further communication on this project, please subscribe to the low-traffic "techlab-announce" e-group at https://e-groups.cern.ch/e-groups/Egroup.do?egroupId=10098252.

CF4Hep Beta Release 0.5

Benedikt Hegner — Fri, 28 Jun 2013 08:27:16 +0000

A first concurrent Gaudi beta release 0.5 was released on Fri 28, 2013.
This is the first step towards enabling the Gaudi event processing framework for running on many-core machines. It enables the framework to

Execute multiple algorithms in parallel
Process multiple events in parallel
Clone of non thread-safe algorithms

A few example workflows in GaudiHive/options show this new functionality.
Together with this release goes an adjusted vertex locator (VELO) local reconstruction of LHCb ("MiniBrunel"), showing the scaling potential of the prototype. This is shown in the plot on this page.
More details about how to install and use concurrent Gaudi v 0.5 can be found at https://twiki.cern.ch/twiki/bin/view/C4Hep/WebHome.

Enjoy!
The Concurrent Gaudi Team

Announcing Annual Concurrency Forum Meeting - February 4-6, 2013

Pere Mato Vila — Mon, 19 Nov 2012 15:58:44 +0000

The Annual Concurrency Forum Meeting will be held on February 4th and 6th, 2013 at Fermilab, with the possibility to overflow parallel discussions a third day if required. This meeting is the follow-up of the first "Workshop on Concurrency in the many-Cores Era" that took place in November 2011. We would like to reserve Wednesday, February 6th for follow up discussions.
During the 2011 Workshop we discussed ideas on how to efficiently exploit future many-core processor architectures and established the "Concurrency Forum" as a communication model to collaborate on this topic. The goals for the February 2013 annual meeting are as follows:
- We would like to discuss progress made with each demonstrator, drawing conclusions for those that have completed their work programme and identifying topics that still need to be investigated.
- It would be nice if each collaboration (experiments, Geant4 etc.) could report on its current views on parallelism issues, such as choice of concurrency model and software technologies. This will be an opportunity for us all to discuss and see if we can converge on a common overall strategy.
- One outcome of the meeting is the possibility of launching development projects with specific deliverables that can be of benefit to the whole community. The extra day can be used by interested groups to get together and to start to discuss a programme of work.
The preliminary agenda and information about travel: https://indico.fnal.gov/conferenceDisplay.py?ovw=True&confId=6138

IT Seminar - Monday, September 24th at 11h

Felice Pantaleo — Wed, 12 Sep 2012 10:29:33 +0000

GPU acceleration of scientific applications: An update
by Dr. Peter Messmer and Mr. Edmondo Orlotti, NVIDIA
Monday, September 24th - IT Auditorium (31-3-004) at 11h
https://indico.cern.ch/conferenceDisplay.py?confId=206360

IT Seminar - Monday, July 23 at 11h

Pere Mato Vila — Fri, 13 Jul 2012 15:18:39 +0000

Dealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-Tuning
by Dr. Victor Pankratius, MIT and Karlsruhe Institute of Technology
Monday, July 23 - BE Auditorium (6-2-024) at 11h
http://cern.ch/Computing.Seminars

WS on Many-core architectures for LHCb

Pere Mato Vila — Fri, 04 May 2012 09:37:57 +0000

LHCb organized a Workshop on many-core architectures for LHCb with main emphasis on ideas, plans and to see what others have done or about to do on this in order to learn from each other. In the context of this concurrency froum there was a presentation on "Adding Concurrency to LHC Software"

Job opening at CERN - PH-SFT-2012-4-LD

Pere Mato Vila — Wed, 18 Jan 2012 15:14:12 +0000

Please note of the publication of a vacancy for the PH-SFT group. This is the post for taking part in the re-engineering of the LHC scientific software used at CERN by introducing parallelism in order to make efficient use of new multi/many-core CPU architectures. Vacancy notice PH-SFT-2012-4-LD.

Mailing lists and archives

Pere Mato Vila — Wed, 18 Jan 2012 14:19:55 +0000

We use the rd-multicore@cern.ch mailing for general announcements and discussion forum.
All mails are archived at https://groups.cern.ch/group/rd-multicore
You can subscribe yourself using a CERN mail account (login required) using the e-groups.cern.ch site.

Announcing first Forum meeting

Pere Mato Vila — Mon, 16 Jan 2012 17:21:49 +0000

We will start regular bi-weekly meetings to share knowledge and discuss progress on a number of 'demonstrators' that were identified in the last Workshop on Concurrency in the Multi-Cores Era help in November. The tine slot is Wednesday at 17:00 (CET) for maximum 1h30. We have booked the room 32-1-A24 for people at CERN and setup a corresponding EVO video meeting for remote participation.
The first meeting will take place this Wednesday 11th of January. We will start with a series of very short presentations from the people who expressed interest to work on the demonstrators. These 2-3 slides presentations should contain information on what their demonstrator is about, what tools and techniques they would be using and what they would like to achieve in 3 months time. Exceptionally, next meeting will be one week later on the January 18th to continue with the short presentations.
The provisional agenda is available at https://indico.cern.ch/conferenceDisplay.py?confId=169634