GlueX Software Meeting, June 11, 2019

From GlueXWiki
Jump to: navigation, search

GlueX Software Meeting
Tuesday, June 11, 2019
3:00 pm EDT
JLab: CEBAF Center A110
BlueJeans: 968 592 007

Agenda

  1. Announcements
    1. New version set: recon-2018_01-ver02_1.xml
    2. MCwrapper: Busy Bee Starsky release
    3. two new version sets: version_4.5.0.xml and recon-2017_01-ver03_10.xml
  2. Review of minutes from the last Software Meeting (all)
  3. Report from the last HDGeant4 Meeting (all)
  4. More on halld_recon, reconstruction launches, and analysis launches (Mark, Thomas)
  5. Review of recent issues and pull requests:
    1. halld_recon
    2. halld_sim
    3. CCDB
    4. RCDB
  6. Review of recent discussion on the GlueX Software Help List (all)
  7. Action Item Review (all)

Slides

Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2019 on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2019/ .

Minutes

Present:

  • CMU: Naomi Jarvis
  • JLab: Alexander Austregesilo, Thomas Britton, Sean Dobbs, Stuart Fegan, Mark Ito (chair), Justin Stevens, Simon Taylor, Beni Zihlmann

There is a recording of this meeting on the BlueJeans site. Use your JLab credentials to access it.

Announcements

  1. New version set: recon-2018_01-ver02_1.xml. Version set for the recon-2018_01-ver02 launch.
  2. MCwrapper: Busy Bee Starsky release. This release includes a new variable ANA_ENVIRONMENT_FILE in MC.config to allow a different version of halld_recon for ROOT tree production than that used for reconstruction in the same job. This follows the ideas in Mark's presentation which came later in the meeting.
  3. two new version sets: version_4.5.0.xml and recon-2017_01-ver03_10.xml. An update release and a 2017-01-reconstruction-launch-compatible upgrade release.

Review of minutes from the last Software Meeting

We went over the minutes of the meeting on May 28 without significant comment.

Report from the last HDGeant4 Meeting

We went over the minutes from the meeting on June 4.

  • Sean asked if the addition of an intermediate buffer when reading EVIO files has been tested on local files, i.e., without use of XROOTD. This can be tested in the current default version of halld_recon.
  • We forgot to note the significant step forward of having an XROOTD server up and running on scosg16.jlab.org (the submit host).
  • Sean mentioned that there may be computing resources at Jülich that we can access in an OSG-like way now that they are collaborators.

More on halld_recon, reconstruction launches, and analysis launches

Mark and Thomas summarized the discussion on the need for separate versions of halld_recon for reconstruction launches and analysis launches. Please see their slides for the full story. The summary slide is reproduced here:

  • Need to identify pairs of halld_recon versions for some tasks
  • Discussed solutions at the last meeting
  • One such solution, two version sets, implemented in MCwrapper
  • Introduced new file, version_set_correlations.xml, to indicate valid pairs of version sets
  • Features:
    • Very little re-coding needed
    • Preserves build_scripts paradigm: a compatible set of tagged releases of software packages
    • Mechanism provided for identifying correlation between reconstruction halld_recon and

analysis halld_recon: new "version set correlation" XML file

Review of recent issues and pull requests

halld_recon issues and pull requests

  • Difference in p 3pi between 2017 low and high intensity data? #160 Sean followed up with Elton after the meeting. This issue is closed now.
  • DL1MCTrigger Crashes With mysql rcdb/ccdb #81 We should follow up with Jon Zarling on this one.
  • crash in ST matching #71 David found and fixed the problem for one offending plugin. See his explanation at the bottom of the issue page. This is a bug of long-standing. It was first reported by Alex on December 18 of last year. Sean and Beni put a lot of work into isolating the problem.
    • Added in press: Sean did a survey of our plugins and corrected the problem in 60 source files in halld_recon.
  • NaN's in the analysis library #52 Sean thought this one might be fixed.
  • Tracking update apr19 #167 Alex showed evidence for small changes in tracking results for a large fraction of tracks as a result of this pull request. He looked at t and proton momentum from recent reconstruction tests. Simon told us that this is not unexpected given that some of these changes cause a different set of hits to be associated with a track. Alex cautioned that there might be some bin migration back and forth when comparing results from before and after this change.

Review of recent discussion on the GlueX Software Help List

We reviewed the list items. Not a lot of comment.