Agenda:

  1. Hodoscope readout                  - Rafo
  2. Readout and Tracking with L0  - Omar
  3. Trigger studies                          - Kyle 
  4. Trigger verification                    - Rafo/Kyle/?
  5. DQM                                         - Matt
    1. We still need some feedback on how many GB per run to write to the /cache disk on their way to tape.  How much is needed for reasonable DQM plots?  Is 2 GB per run enough?  10 GB per run?
  6. AOB/Questions

Notes:

Rafo - Hodoscope readout.

  • First part: EVIO to LCIO is done and merged.
  • Second part: 
    • Decided to use generic collections, because the data is different from Ecal. 
    • Hodoscope Hits: Several arrays: x-index, y-index, layer and energy, and (question).
    • Hodoscope Clusters: Combines the hit that are in the same tile, that have dual readout.
    • Used the detector with only hodoscope and ecal and it worked. With other detector, there seems to be an issue with the conditions?

 

The Readout, Tracking and L0 issue (Norman):

  • Norman: Matt Solt made some changes. (Omar is not online, Matt is in an airplane).
  • They are working on it. Norman hopes to get some actual data. It appears to work (they find hits and tracks, and hits are added to the tracks in the simulation reconstruction.)
  • The hits show up from MC and are fully analyzed and tracks are found. 
  • We need to look at some data soon!

 

Trigger (Kyle)

  • A' MC is available now. Readout completed on this data. The recon did not produce tracks because of a timing issue with the tracks.
  • Today the "truth" distributions should complete. This will give us the final cuts.
  • We will also need a full background simulation for getting the rates. 
    • Bradley has some issues with the new batch system, SLURM. (Seems to use bash instead of tcsh?) 

Trigger Verification:

  • This needs tracking to work. 
    • Check if track points to ECAL, and then check the trigger banks.
    • The trigger objects are now ready.
    • Some very low threshold noise data was used to check the decoding. This works.
    • The VTP conversion to generic object is merged.
    • Verification: Convert LCIO to ROOT NTuple and then run analysis over the output in ROOT.

DQM (Matt)

  • Q: How much data is needed to do a reasonable DQM?  
    • A single data file for 2016 V0 skim. 12k events in the file, 38k clusters, 27k tracks, ~ 5k tridents. From a single 2GB data file.
    • Occupancies, etc are also available in online monitoring.
    • How many tridents do we need? Not clear.
    • We should try to do at least 1% of data for DQM, perhaps 2-5%?

Questions: 

  • Norman to Rafo: What driver for the hodoscope?   
    • Rafo will post the reconstruction driver info. It needs to be added to the standard steering file.
  • Kyle: For the trigger studies, needs the correct method for extending a track to the ECAL face. There isn't a cluster-track matching for the GTP clusters.
    • Norman will show Kyle where the files are and how to get going.

 

 

 

  • No labels