06/18/2013 Notes

DocDB Event


Jan, Zukai, Matthew, Will, Andrew, Alec, Barnali and Martin (taking notes)

Trigger Message

  • The partition number is configurable using the FHICL file.
  • It turns out that passing a command line argument into ddt-filter is a bit difficult:
    • The DDT-specific parameter needs to be stripped out of argc and argv before passing them on to the ART application.
    • A ParameterSet cannot be modified, you have to go through the Intermediate Table first.
    • For now, let's live with the FHICL parameter.
    • Andrew suggests that we can also read the partition out of the environment. However, this requires that we set it first in the DDT setup script. Only DAQ sets the variable so far.
  • DAQ run with DDT message on Thursday morning 9 - noon with Andrew and Alec. Andrew will let Peter know as well.


  • Matthew gave his talk which was very nice!
  • Andrew explains that the raw data arrives grouped by DCM, but not sorted by DCM. Within each DCM block the data is roughly sorted by the arrival time.
  • The simulation does not have this ordering, so TDC sort might be faster on the raw data instead of the simulation.
  • Why does the hitProducer take so much time? For simulation, the hitProducer copies the data from the BlueArc over the network which is a very small pipe. We believe that it is mainly due to IO issues and we expect it to be faster for running on the raw data.
  • I will send Matthew the information that I found a while back.


  • Jan showed his talk which was also very nice, yet there is some work left to be done.

Zukai's AWESOME Progress

  • He made wonderful progress that I skipped over, so we did not have the pleasure of hearing it.
  • We'll hear about it in two weeks then!


  • Zukai and I will be out next week and Andrew will be out Friday and Monday.
  • Barnali joined us for the first time from South Carolina. She will be at Fermilab in two weeks.