Developer's meeting

→ Europe/Zurich

present: Milo, Enrico, Frank, Kai, Junbin, Eric

EHN1 integration:

  • f/w: started an effort (together with Andrea and Frans) to help throughput:
    • more DMA descriptors (one per thread): dual descriptor working on VC709 while still some stability issues on BNL711
    • bigger blocks (to reduce memory copies): version exists that can deal with 9 complete chunks (needs work on stability)
  • s/w: Joern provided a lightweight felixcore version (should work with dual DMA)

BNL investigation of data losses:
(see slides)

  • no issues seen with "ATLAS" Full mode, or when using internal emulation mode; strong suspicion that the issue is with the WIB input
  • most recent WIB f/w version may not have been used here, so it would be worth updating to that
  • concluded that Dan is not using the 8b/10b encoder firmware block (as provided by the FELIX team) in the way expected; iterating with him

Update on BNL WIB-FELIX testing:
(see slides)

  • fixed channel mapping issues
  • while pulse injection works with LabView, it does not seem to do so when using BUTools
  • continuing work on configuration and analysis scripts

Next meeting: tentatively planned for April 25

There are minutes attached to this event. Show them.