GlueX Offline Meeting, October 15, 2014

From GlueXWiki
Jump to: navigation, search

GlueX Offline Software Meeting
Wednesday, October 15, 2014
1:30 pm EDT
JLab: CEBAF Center F326/327


  1. Announcements
    1. EVIO support in Public Builds
  2. Review of minutes from September 17 (all)
  3. Review of Talks from Collaboration Meeting
  4. commissioning simulations
  5. offline analysis of online data and coordination thereof
  6. Action Item Review

Communication Information

Remote Connection


Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2014-4Q on the JLab CUE. This directory is accessible from the web at .



  • CMU: Paul Mattione, Curtis Meyer
  • IU: Matt Shepherd
  • JLab: Mark Ito (chair), Manuel Lara, David Lawrence, Will McGinley, Kei Moriya, Eric Pooser, Sascha Somov, Nathan Sparks, Mike Staib, Simon Taylor, Aristeidis Tsaris, Beni Zihlmann
  • MIT: Justin Stevens
  • NU: Sean Dobbs
  • UConn: Richard Jones


  • EVIO support in Public Builds. Mark has added EVIO-support to the nightly builds and the most recent release (2014-09-23). See his email for more details.
  • Disk location of nightly builds at JLab. Mark moved the nightly builds at JLab from the group disk to the scratch disk. See his email for details.

Review of Minutes from September 17

We looked over the minutes.

Sean has implemented the Jones-Dobbs scheme we discussed for solving the 100-ns-Shift/effective-time-cut issue that arises when converting HDGeant's HDDM output to EVIO format. Time offsets are detector dependent. One must use the "mc" CCDB variation when reconstructing this kind of EVIO-formatted data to get the right constants.

Commissioning Simulations

Sean reported that jobs were run over the weekend, 2 k jobs of 30 k events each (1 k of the jobs with field, 1 k without). The success rate has not been measured. The cause of many crashes is a problem in the BCAL smearing routine. David and Sean are looking at the problem. Despite this the data should be usable.

Richard and Justin pointed out that events with hadronic events are not useful for rate studies. Sean will put together a configuration for EM background only runs so we can get started. Richard also cautioned us that the way we are now generating the incoherent photon spectrum may change the normalization calculation vis-a-vis the way it is obtained when generating the standard coherent spectrum. Sean will discuss this with Richard offline.

Richard is ready to include Pavel Detiarenko's physics package into the tagger hall simulation. Mark agreed to contact Pavel about when it will be ready.

Offline Data Quality Assurance

Paul moderated the discussion.

Online Monitoring Histograms

Kei summarized the status of the online monitoring plugins in the following table:

Detector developing? test with data? notes
FCAL YES YES Matt, Manuel working on plugin
PSC NO NO MEPHI students?
PS NO NO MEPHI students?
ST YES YES Eric working on plugin
TAGH YES NO? Nathan working on plugin
TAGM NO? NO? Richard?
TOF YES NO? Alexander, Aristeidis working

Matt asked how histogram/macro selections will work, since not all histograms generated in the plugin will be of general interest. David reminded us that the ROOTSpy GUI allows the user to select the histograms and/or macros results that should be displayed. Different detectors are selected on separate tabs in the GUI. There will be a default selection done, but private custom selection sets are supported. To make a macro available for selection, it needs to be checked into the same directory as the plugin code.

Monitoring Database

Sean has a first draft database ready to go. He has installed it on the Hall D offline MySQL server, He is working on the scripts to fill it from processing jobs. Mark remarked that his database is a candidate for replication to the counting house.

Run Browser

Justin reminded us of the location of the Run Browser and Time Series webpages. Results of the monitoring effort will appear there.


David is looking at using Gagik Gavalian's EVIO-to-ROOT tool. I would be hard to adopt for our current raw data format, but he thought that if we re-write the data, as will be necessary for blocked data, we might have an output EVIO format that would match the tool's expectations.

Action Items

  1. Figure out the incoherent bremsstrahlung spectrum normalization. (Sean, Richard)
  2. Give a talk on a software versioning scheme at next week's Calibration Meeting. (Mark)