US ATLAS Software and Computing Technical Planning Meeting 2015

→ Europe/Zurich
40/S2-D01 - Salle Dirac (CERN)

40/S2-D01 - Salle Dirac

CERN

95
Show room on map
Mark Neubauer (Univ. Illinois at Urbana-Champaign (US)), Michael Ernst (Unknown), Nurcan Ozturk (University of Texas at Arlington (US)), Torre Wenaus (Brookhaven National Laboratory (US))
Description

This meeting combines

  • one in the regular series of facilities/distributed software F2F meetings
  • the annual US ATLAS software technical planning meeting

Developers and technical experts are invited to this meeting so that it can be a substantive technical discussion and planning session for key US activities. Towards making it substantive and technical, the topics will be limited.

All ATLAS members interested to attend are welcome.

A follow-up management meeting at Duke in July will factor the outcomes of this meeting into the planning for the next year, and will scrub the S&C and Physics Support programs.

Sorry about the pin, it is 0628

Meeting notes - editable googledoc -- please contribute

    • 10:00 → 13:15
      Core services, event store
      • MT framework and the US role, common efforts (CMS?), US role in Gaudi(Hive), G4Hive
      • where next for athenaMP
      • breaking G4 simu away from athena, statically linked G4 with performance improvement, G4MT
      • athena platform porting (powerPC...)
      • where next for event I/O, in-file metadata, US role
      • where next for ROOT6
      • event streaming service: specific objectives, workplan and timeline
      • US planning issues for the rel 21 schedule
      • longer term program to run 4
      Conveners: David Malon (Argonne National Laboratory (US)), Nurcan Ozturk (University of Texas at Arlington (US)), Paolo Calafiura (Lawrence Berkeley National Lab. (US))
      • 10:00
        Future Framework 40m
        Speaker: Charles Leggett (Lawrence Berkeley National Lab. (US))
        Slides
      • 10:40
        Multiprocessing in Run 2 and beyond 20m
        Speaker: Vakho Tsulaia (Lawrence Berkeley National Lab. (US))
        Slides
      • 11:00
        Machine Learning and Run 4 Computing 15m
        Speaker: Paolo Calafiura (Lawrence Berkeley National Lab. (US))
        Slides
      • 11:15
        Coffee Break 15m
      • 11:30
        Toward a next-generation I/O framework 45m
        integrated into ATLAS future framework, removing the last vestiges of POOL/APR, supporting mulithreading and distributed processing, suitable for HPC deployment
        Speakers: Alexandre Vaniachine (ANL), David Malon (Argonne National Laboratory (US)), Jack Cranshaw (Argonne National Laboratory (US)), Marcin Nowak (Brookhaven National Laboratory (US)), Peter Van Gemmeren (Argonne National Laboratory (US))
        A framework-oriented bridge to metadata discussions
        Toward a next-generation I/O framework
      • 12:15
        The future of metadata and bookkeeping infrastructure 45m
        why it matters, what the issues are, near-term and longer-term work
        Speakers: Alexandre Vaniachine (ANL), David Malon (Argonne National Laboratory (US)), Jack Cranshaw (Argonne National Laboratory (US)), Marcin Nowak (Brookhaven National Laboratory (US)), Peter Van Gemmeren (Argonne National Laboratory (US))
        File Peeking
        Post Reco
    • 14:00 → 16:15
      Distributed software, HPC
      • where next for PanDA/JEDI, Prodsys2, monitoring, UberPanDA
      • collaborative work on UberPanDA (=catch-all for Big, Mega, Exa etc), benefits to/relation with ATLAS work
      • plans for new developments, psub, distributed aspects of event streaming service, data product catalog, ...
      • US HPCs, ops program roles, relation to non-ops & other ATLAS work, convergence. Objectives, timeline for US HPC prodsys production
      Conveners: Alexei Klimentov (Brookhaven National Laboratory (US)), Kaushik De (University of Texas at Arlington (US)), Nurcan Ozturk (University of Texas at Arlington (US))
      • 14:00
        Introduction and outline of the session 10m
        Speakers: Alexei Klimentov (Brookhaven National Laboratory (US)), Kaushik De (University of Texas at Arlington (US))
        Slides
      • 14:10
        Distributed Software 55m
        Speakers: Alden Stradling (University of Texas at Arlington (US)), Alexandre Vaniachine (ATLAS), Paul Nilsson (Brookhaven National Laboratory (US)), Sasha Vanyashin (Argonne National Laboratory (US)), Siarhei Padolski (BNL), Tadashi Maeno (Brookhaven National Laboratory (US))
        • ProdSys2 : 1 year progress and plans. Where we are ? What is missing ? What are our priorities ? 15m
          Speaker: Tadashi Maeno (Brookhaven National Laboratory (US))
          Slides
        • PanDA pilot : progress, plan, future evolution 15m
          Speaker: Paul Nilsson (Brookhaven National Laboratory (US))
          Slides
        • ProdSys2 and PanDA monitoring 10m
          Speaker: Siarhei Padolski (BNL)
          Slides
        • Contribution to ADC SW (fraction of FTE experience) 10m
          Speakers: Alden Stradling (University of Texas at Arlington (US)), Alexandre Vaniachine (ATLAS)
          Slides
      • 15:05
        Distributed SW Q&A. Discussion 10m
      • 15:15
        HPC&LCF 40m
        Speakers: Danila Oleynik (Joint Inst. for Nuclear Research (RU)), Sergey Panitkin (Brookhaven National Laboratory (US)), Taylor Childers (Argonne National Laboratory (US)), Vakho Tsulaia (Lawrence Berkeley National Lab. (US))
        • Titan : ATLAS and non-ATLAS payloads, integration with ATLAS, pilot for HPC, progress, issues, future 15m
          Speakers: Danila Oleynik (Joint Inst. for Nuclear Research (RU)), Sergey Panitkin (Brookhaven National Laboratory (US))
          Slides
        • MIRA : Progress, issues, plans, integration with ATLAS Production, How to address physics priorities 10m
          Speaker: Taylor Childers (Argonne National Laboratory (US))
          Slides
        • Event Service, Yoda, NERSC : progress, problems, plans 15m
          Speaker: Vakho Tsulaia (Lawrence Berkeley National Lab. (US))
          Slides
      • 15:55
        HPC&LCF Q&A 10m
    • 16:00 → 16:30
      Software Infrastructure
      Convener: Alexander Undrus (Brookhaven National Laboratory (US))
      • 16:00
        ATLAS software infrastructure: requirements and goals at Run 2 period 15m
        Speaker: Alexander Undrus (Brookhaven National Laboratory (US))
        Slides
      • 16:15
        Challenges of ATLAS software environment setup 15m
        Speaker: Shuwei Ye (Brookhaven National Laboratory (US))
        Slides
    • 16:30 → 18:30
      Tier 3 Round 1
      • the needs, the plan, the technologies and approaches, the ops program role
      • LOCALGROUPDISK policy in support of US analysis
      • talk things through to the point where we're able to write down an ops program plan
      Conveners: Doug Benjamin (Duke University (US)), Erich Ward Varnes (University of Arizona (US)), Michael Ernst (Brookhaven National Laboratory (US))
    • 08:25 → 11:30
      US ATLAS facilities, distributed computing
      • integrated facility issues & plans for supporting Run 2+ computing
      • what's next for FAX, xrootd, federated data access, wide area data access
        • a cross-cutting topic that needs to involve core sw, dist sw also
      • what's next for APF
      • plans & expectations for opportunistic platforms: Amazon spot market, OSG, ...
      Conveners: Michael Ernst (Brookhaven National Laboratory (US)), Robert William Gardner Jr (University of Chicago (US))
      • 08:25
        Introduction and Session Outline 5m
        Speaker: Michael Ernst (Brookhaven National Laboratory (US))
      • 08:30
        What's next for FAX, brokering, federated wide area data access? 30m
        Speakers: Ilija Vukotic (University of Chicago (US)), Kaushik De (University of Texas at Arlington (US))
        Slides
      • 09:00
        Client distributed IO performance discussion: framework, ROOT, xrootd, ... 20m
        Speakers: Andrew Hanushevsky (STANFORD LINEAR ACCELERATOR CENTER), David Malon (Argonne National Laboratory (US)), Ilija Vukotic (University of Chicago (US))
        Slides
      • 09:20
        APF News and Plans & Expectations for Opportunistic Platforms 20m
        Speakers: John Raymond Hover (Brookhaven National Laboratory (US)), Jose Caballero Bejar (Brookhaven National Laboratory (US))
        Slides
      • 09:40
        Plans & expectations for opportunistic platforms: OSG, ATLAS Connect 20m
        Speaker: Robert William Gardner Jr (University of Chicago (US))
        Slides
      • 10:00
        Xrootd Caching 20m
        Speakers: Ilija Vukotic (University of Chicago (US)), Wei Yang (SLAC National Accelerator Laboratory (US))
        Slides
      • 10:20
        SDN Provisioning, next steps after ANSE 20m
        Speaker: Kaushik De (University of Texas at Arlington (US))
        Slides
    • 11:30 → 12:30
      Tier 3 round 2
      • write down and discuss a specific ops program plan
      Conveners: Doug Benjamin (Duke University (US)), Erich Ward Varnes (University of Arizona (US)), Michael Ernst (Brookhaven National Laboratory (US))