Computing R&D Workshop 2011

Europe/Rome
Ferrara

Ferrara

IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
    • Introduction, SuperB Status and goals for the workshop IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy

      EVO meeting:
      http://evo.caltech.edu/evoNext/koala.jnlp?meeting=MtM8Ma2i22D8De9u9sDs9t

      • 1
        Welcome
        Speaker: Diego Bettoni (FE)
      • 2
        Status of SuperB
        Speaker: Fabrizio Bianchi (TO)
        Slides
      • 3
        Goals of the Workshop
        Speaker: Dr Peter Elmer (Princeton University)
    • 16:30
      Coffee break
    • General Computing Requirements IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
      • 4
        Computing Resource Estimates for SuperB
        Presentation of SuperB computing resource estimates based on an extrapolation from the BaBar computing model. Added 20.9.2011 (Steffen): Please note that cost numbers in the talk are incorrect (based on a typo in the spreadsheet). Please refer to the updated spreadsheet for the correct numbers.
        Speaker: Dr Steffen Luitz (SLAC)
        Slides
        Spreadsheet
      • 5
        CMS Computing Resource Planning
        The resource planning being done in CMS Computing is taken as an example. The experience in designing, building and running a monthly-granular computing resource spreadsheet is presented. Differences as well as common points with Super-B are highlighted and discussed.
        Speaker: Daniele Bonacorsi (BO)
        Slides
      • 6
        Computing Requirements and Crystal Balls
        Wild guesses of what might happen in industry and SuperB computing over the next 5-10 years.
        Speaker: Davide Salomoni (CNAF)
        Slides
      • 7
        Discussion
    • New CPU Architectures, GPU's and Framework (I) IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy

      EVO meeting:
      http://evo.caltech.edu/evoNext/koala.jnlp?meeting=eBele2vtv9anaeIeauIB

      • 8
        HEP software on Multicore: present experience and perspectives
        Speaker: Vincenzo Innocente (CERN)
        Slides
      • 9
        A Parallelization Exercise
        Speaker: Stefano Longo (PD)
        Slides
      • 10
        Performances of some HEP applications
        Speaker: Vincenzo Ciaschini (CNAF)
        Slides
    • 10:30
      Coffee break
    • New CPU Architectures, GPU's and Framework (II) IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
      • 11
        Event Processing Frameworks
        Speaker: Dr Peter Elmer (Princeton University)
        Slides
      • 12
        Physics Simulations on multi- and many-core architectures
        Speaker: Sebastiano Fabio Schifano (FE)
        Slides
    • 13:00
      Lunch
    • Machine Control System/ETD/Online IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
      • 13
        Accelerator Computing Infrastructure & Controls R&D Introduction
        The SuperB Accelerator complex requires a large amount of computing tools mainly dedicated to three different purpose: implementation and maintenance of documentation and project management; beam simulation and controls; data monitors, presentation and correlation with the experiment. In the mean time those tools require identification, security, accessibility for large and international community. This is an opportunity to develop new concepts in accelerators controls and realize a new computing infrastructure. Requirements, preliminary ideas and structure design will be presented.
        Speaker: Giovanni Mazzitelli (LNF)
        Slides
      • 14
        The SuperB Accelerator Control System - Plans and R&D Status
        !CHAOS is a new project under development at INFN aimed at the design and validation of a new paradigm of control system for particle accelerators that goes beyond the so called standard model. The key points of the new development are the high level of abstraction of services and components, the implementation of high-performance software technologies for the continuous distribution and storage of data, a new topology allowing high scalability of services and performance, an intrinsic redundancy of components avoiding any point of failure. The talk will present the !CHAOS design concept, the preliminary results, the future research program and the plans in the perspective of SuperB project.
        Speaker: Luciano Catani (ROMA2)
        Slides
      • 15
        ETD/Online for the SuperB Detector - Plans and R&D Status
        This talk will give an overview of the proposed Electronics, Trigger, Data Acquisition and Online (ETD/Online) system for the SuperB detector and the associated R&D activities. The current baseline design will allow up to 150 kTriggers/s and 75kByte event size with negligible dead time (<=1%). It builds upon lessons learned in the BaBar and LHC experiments and implements simple, fully synchronous front-ends, a scalable ROM and back-end architecture and utilizes commercial off-the-shelf components where possible.
        Speaker: Dr Steffen Luitz (SLAC)
        Slides
      • 16
        (databases)
        Speaker: Luca Tomassetti (FE)
        Slides
      • 17
        Discussion
        Open discussion of areas of joint interest, common R&D, etc.
    • 16:30
      Coffee break
    • Software Development Tools IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
      • 18
        Quality check tools for software developers
        Speaker: Francesco Giacomini (CNAF)
        Slides
      • 19
        ALICE experience using Coverity
        This talk will convey the experience of using Coverity for AliRoot software in A Large Ion Collider Experiment (ALICE). AliRoot consistently undergoes intensive development and release cycles, thus creating an increasing demand for automated high precision tools aiding developers in maintaining coding standards and ensuring code correctness. Coverity as a software solution, providing both static code analysis facilities and a centralised web interface, is actively enforced within the collaboration. The talk will start by describing the approach taken to set up Coverity and the automation process. It will also look at the use policy and the observed effects. The presentation will then explore checkers of varying impact with few examples of observed defects. In addition, examples of false positive cases will be presented with an explanation on how to model Coverity to ignore them. Thirdly, the talk will provide a critical evaluation of the application of Coverity to the AliRoot code with some considerations on the usage of automated tools in High Energy Physics (HEP) experiment software development. In conclusion, the talk will explore additional tools extensively used within ALICE for ensuring good coding standards.
        Speaker: Olga Datskova (Cern)
        Slides
      • 20
        Experience in using Etics, a multi-platform and open source build and test system for big software projects
        Speaker: Elisabetta Ronchieri (Cnaf)
        Slides
      • 21
        CMS build and integration workflows: present and future perspectives
        Speaker: Giulio Eulisse (Northeastern Uni)
        Slides
      • 22
        Evaluation of the Git Version Control System
        Speaker: Marco Corvo (Cnrs & INFN)
        Slides
    • Distributed Computing (I) IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy

      EVO meeting:
      http://evo.caltech.edu/evoNext/koala.jnlp?meeting=eBele2vtv9anaeIeauIB

      • 23
        Future of distributed data management systems
        Speaker: Dr Fabrizio Furano (CERN)
        Slides
      • 24
        Future of Grid parallel exploitation, Grid MPI
        Parallelism support in the current grid middleware (gLite) and in the forthcoming EMI middleware will be overviewed, with special focus on the new features (support of the Wholenodes reservation, CPU affinity and hybrid MPI-openMP programming). Eventually I will shortly present the experience of parallel clusters installed in grid and supporting a preliminary version of these new features.
        Speaker: Roberto Alfieri (PR)
        Slides
      • 25
        Future of Grid resource exploitation systems
        Speaker: Ricardo Graciani (Barcellona University)
        Slides
      • 26
        Discussion
    • 10:30
      Coffee break
    • Distributed Computing (II) IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
      • 27
        LHC current and future proposals
        In this talk we'll present the current solutions adopted for the LHC Computing and their evolutions, with a final overview of a possible future scenario of the HEP Computing Models and tools.
        Speaker: Alessandro De Salvo (ROMA1)
        Slides
      • 28
        Evolution of the distributed computing model - The case of CMS
        The presentation starts from the description of how the CMS experiment is changing its approach in the exploitation of distributed resources and concludes with thoughts on possible evolutions for the future of HEP experiments.
        Speaker: Claudio Grandi (BO)
        Slides
      • 29
        Middleware Evolution in EMI
        Speaker: Dr Francesco Giacomini (CNAF)
        Slides
      • 30
        Discussion
    • 13:00
      Lunch
    • Storage (I) IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
      • 31
        SuperB R&D on going on storage and data access
        Speaker: Dr Silvio Pardi (INFN Napoli)
        Slides
      • 32
        The Storage @ INFN Tier1: status and perspective
        Speaker: Luca Dell'agnello (CNAF)
        Slides
      • 33
        Distributed Tier1: NDGF experience and future developments
        Speaker: Gerd Behrmann (NDGF)
        Slides
      • 34
        EMI Storage Evolution
        Speaker: Patrick Fuhrmann (DESY)
        Slides
      • 35
        HDFS and Remote Xrootd in a CMS Tier2
        Speaker: Dr Brian Bockelman
        Slides
      • 36
        Discussion
        Speaker: All
    • 16:30
      Coffee break
    • Storage (II) IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
      • 37
        DataManagement in LHC experiments
        Speaker: Daniele Bonacorsi (BO)
        Slides
      • 38
        Experience and future prospects of Lustre at a large Tier 2 site
        Speaker: Alex Martin (Queen Mary, University of London)
        Slides
      • 39
        Discussion
        Speaker: All
    • R&D support platforms & Action Items Discussion IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy

      EVO meeting:
      http://evo.caltech.edu/evoNext/koala.jnlp?meeting=MIM9Ms2D2nDDDv9v9tDa9I

      • 40
        Planning of SuperB Computing Infrastructure in Southern Italy
        Speaker: Guido Russo (NA)
        Slides
      • 41
        Action Items from R&D on new computing architectures
        Speaker: Vincenzo Innocente (CERN)
        Slides
      • 42
        Summary from Accelerator Controls / ETD/Online Session and Action Items
        Speaker: Dr Steffen Luitz (SLAC)
        Slides
      • 43
        Action Items on Software Development Tools
        Speaker: Mr Marco Corvo (PD)
        Slides
    • 10:30
      Coffee break
    • Action Item discussion and Closeout IUSS

      IUSS

      Ferrara

      IUSS Ferrara 1391 via Scienze 41b 44100 Ferrara, Italy
      • 44
        Action items on Distributed Computing
        Speaker: Armando Fella (PI)
        Slides
      • 45
        Action Items on Storage
        Speaker: Dr Silvio Pardi (NA)
        Slides
    • 13:00
      Lunch