Developer's meeting

→ Europe/Zurich

Present: Enrico, Roland, Giovanna, Eric, Junbin, Frank

EHN1:

  • f/w (Enrico):
    • now looking into multi-chunk solution, #chunks=2 working
      • this leads to CRC errors but this is understandable, given that (as a quick fix) only the CRC of the last chunk in a multi-chunk is checked
    • now finishing version with #chunks=6 (but there is no intrinsic limitation to this number)
  • BoardReader (Giovanna, Roland):
    • now have emulator (nonsensical data but packaged as artdaq fragments; can be configured for use in .fcl)
    • investigation (in parallel to multi-DMA development) of Jos's alternative based on scatter/gather techniques:
      • need to pin threads to CPUs
      • but even then the CPU resources required for the networking appear to be excessive
    • continuing to investigate Infiniband as alternative

IEEE proceedings for Milo's poster (to-do items before sending to DUNE APB):

  • ADC noise: came to the conclusion [after the meeting] that it may be appropriate to quote the uBooNE paper for the 4 ADC counts noise figure
  • Giovanna: actual references not needed for hardware or software
  • there are some missing numbers... it is not evident that we will be able to fill these in, so we may need to be more vague than necessary

AOB: not actually discussed during the meeting, but the proposal is to meet next in two weeks' time, as usual (the following week will not work for some)

There are minutes attached to this event. Show them.
    • 17:00 → 17:20
      EHN1 Integration 20m
      Speakers: Enrico Gamberini (CERN), Roland Sipos (CERN)
    • 17:20 → 17:40
      IEEE proceedings 20m