Project

General

Profile

04/09/2013 Notes

Attendance

Andrew, Zukai, Brian, Matthew, Yuri, Jan, Louise, Alec, and Martin (taking notes)

General

  • Andrew will request working time not opposite of Exotics for the collaboration meeting.
  • No DDT Meeting next Tuesday!

DDS Messaging

  • Gennadiy helped set up all of the necessary UPS products.
  • I was able to get the FD setup finished.
  • The current issue is that the rms library is not position independent, so Gennadiy showed me how to compile it with the -fPIC option.
  • Andrew talked with Kurt and they are surprised that it is not already compiled with -fPIC. They want me to bring this up at the DAQ Meeting this week.
  • If I can get the messaging working, we would like to try sending a message to the Global Trigger:
    • Andrew suggests to first send messages manually (outside of DDT).
    • Alec and Adam have a test executable for that.
    • Andrew and Alec will talk on Friday afternoon.
    • Alec mentions that there is a DAQ software wiki about the global trigger.
    • Alec will put a wiki entry together for DDT, so I can start toying around before Friday.
  • Zukai and I were able to generate 5 ms FD cosmic MC events:
    • I set MockDataDAQ delay = 0, so all the events now start at 0.
    • Zukai was able to increase the generation window to 5 ms.
    • Zukai notes that there is some limitation on how many 5 ms seconds will fit into one MC file.
    • Brian notes that you can set the number of events per file to anything you want.
    • We have 100 cosmic files with 5 events of 5ms time window (~100 MB)
    • Zukai was able to convert these files using MockDataDAQ and then generate a DDT Root file from that.

MC Files

  • Brian and Jan have previously requested that there be less files when generating MC.
  • Zukai suggested last week that this can be achieved by running the DDT framework over the RAW files without any higher level modules and write them out.
  • Brian then stated that this does not get rid of his 200 RAW files and that he would prefer an automatic management of the files.
  • Andrew suggests a script that will take care of all these steps: running MockDataDAQ, running DDT and then remove the intermediary RAW files.

PatRec

  • Zukai added two new members to the track object: fView (0 = yz, 1 = xz) and fInverseBeta (inverse of velocity: time / distance)
  • Tracks are 2D objects.
  • Brian asked Zukai to update his track fitter to work with the new Track object.
  • Zukai also changed HoughPoint to add some scaling to account for the difference in cell width as compared to plane width.
  • Zukai found some logic flaw in his code. It now runs over the 5 ms cosmic event in ~1.5 second (as compared to ~22 s at the beginning of the week).
  • Zukai also took a look at the supernova trigger for Alec (DocDB 8912).

PackageVersion

  • We will now use PackageVersion from the DAQ and not the UPS product anymore.
  • I will make these changes.

DDT Manager

  • Kurt and Jon are currently working on the resource manager.
  • Andrew suggests that we should just make DDT another resource.

Jan's Update

  • Jan was able to run over the MC and populate histograms with his analyzer module.
  • Jan attempted to install the DDT framework at Minnesota.
  • Jan needs an Online tag urgently. Andrew will talk to Peter about it.

Code Optimization

  • When I ran Zukai's reconstruction module at the far detector it took ~1.51 s using the "prof" binaries and ~1.48 s using the "debug" binaries.
  • It appears that the "prof" compilation does not magically accelerate the code.
  • Andrew suggests that I talk with Marc Paterno about the proper optimization flags for us.