Go to the Indico Home Page Oldest event Older event Up to category Newer event Newest event Printable version
  • Compact style
  • Indico style
  • Indico style - inline minutes
  • Indico style - numbered
  • Indico style - numbered + minutes
  • Indico Weeks View
    • Switch to the management area of this event
    Login

    Developer's meeting

    → Europe/Zurich
    Hide

    present: Roland, Giovanna, Eric, Milo, Frank, Junbin, Kai, Brett

    EHN1 integration:

    • after BR implementation: continued focus on throughput optimisation
    • now three versions:
      • FDR version by Jörn: i/o vectors => 5.7 Gb/s max
      • optimisation using Infiniband support: Infiniband w/ single buffer: 20 Gb/s (single full-mode link) -- but unable to use this from within felixcore
      • publishing through fdaq (using Infiniband): 7 Gb/s, uses ~ 70% of one core (the reduction w.r.t. the above 20 Gb/s is thought to be solvable)
    • next step: attempt to use in felixcore
    • felixcore: extraction of chunks uses a lot of CPU
    • NB: new WIB f/w available, also needs to be tested

    BNL WIB-FELIX testing:

    • (see slides)
    • now testing with one WIB, 4 FEMBs (256 channels); different pedestal settings for induction/collection wires
    • see different noise for different gains (seems to scale with gain)
    • using fdaq/fcheck, see data errors for data sent by WIB (e.g missing data) but not when using internal emulation
    • also observe some issues by data sent from felixcore to subscriber (testBW) when using FLX712


    DUNE R&D plans:

    • (see slides; these were shown by Brett also the preceding day, in the DUNE DAQ PI meeting)
    • two FELIX oriented R&D tasks were suggested:
    • demonstrate feasibility of using FELIX to receive dual-phase data from uTCA crates
      • DP aims to use UDP (data are sent by commercial modules mounted on uTCA crates. It does not seem straightforward to change this
        Giovanna's suggestion: BNL is best equipped to study if alternatively, FELIX can adapt to UDP input data
      • additional item (but less FELIX specific): DP trigger primitive finding
    • discussion: it would be good to see who could contribute what to these studies
      • [ note: the above is in addition to already existing SP tasks ]

    Next meeting: tentatively on April 18

    There are minutes attached to this event. Show them.
      • 17:00 → 17:20
        EHN1 progress 20m
        Speakers: Enrico Gamberini (CERN), Roland Sipos (Institut of theoretical physics Eotvos Lorand University (HU))
      • 17:20 → 17:40
        WIB-FELIX tests at BNL 20m
        Speaker: Junbin Zhang (Brookhaven National Laboratory (US))
        WIB-FELIXmeeting-040418-v2a.pdf
      • 17:40 → 18:00
        FELIX oriented DUNE DAQ R&D plan 20m
        Speaker: Brett Viren (Brookhaven National Laboratory)
        tasks.pdf