[Halld-offline] Minutes, GlueX Offline Meeting, June 15, 2010

[Mark M. Ito]

Find the minutes of the June 15, 2010 GlueX Offline meeting posted at 
http://www.jlab.org/Hall-D/software/wiki/index.php/GlueX_Offline_Meeting%2C_June_15%2C_2010#Minutes 
and reproduced as text below.

GlueX Offline Meeting, June 15, 2010
Minutes

    JLab: Mark Ito (chair), David Lawrence, Sascha Somov, Simon Taylor,
    Elliott Wolin, Beni Zihlmann

Announcements

     1. We looked at the new [29]Sim-Recon Tagged Releases wiki page.
     2. We reviewed Richard's recent wiki page on [30]How HDGeant defines
        time-zero for physics events.


Review of minutes from the last meeting

    We went over [32]them. No significant comments.

GlueX Software Workshop

    David told us that he has called an organizational meeting on Thursday.
    Ole Hansen, Ed Brash and David Heddle will attend. Simon, Beni, Sascha
    may also show up.

Bug/issue tracking software

    The system is up and seems to be working. Elliott solicited feedback
    from non-JLab collaborators and got some: system looks OK, but there
    are too many emails. Elliott explained that that was kind of on purpose
    for the test.

    We will deploy the "real" version, with a less verbose email
    configuration, for use and testing among a small group.

Test of JANA on a 48-core machine

    David showed some nice results he got from running multi-threaded
    analysis code on a demo unit that the Lab has obtained recently. See
    [36]his slides for details. Scaling of performance with large number of
    cores looks very, very promising. Some versions of the analysis code
    seem to scale better than others in this initial look. More work is
    needed to understand the behavior.

Standard math library, matrices, vectors, etc.

    We discussed whether there was a need to pick a standard package for
    standard mathematical objects, much like CERNLIB in the old days. Mark
    has pointed out that we have already de-facto picked ROOT as the
    provider of these services, despite the fact that typedef'ed proxies
    are actually referenced in the code. CLHEP is another candidate for
    this function; it is used by GEANT4. The consensus was that an official
    stance on the issue was not necessary. As code that uses other packages
    (CLHEP in particular) is introduced, then the new service packages can
    be required as auxiliary software if the case is compelling.

New Action Items

     1. Add comments to sim-recon release page. -> Mark
     2. Announce field-uniformity GlueX note. -> Simon
     3. Give a plug-in tutorial to the group -> David
     4. Release JANA 0.6.2 -> David
     5. Put old action items into MantisBT. -> Sascha
     6. Bring up DAQ software release method at an online meeting. ->
        Elliott
     7. Configure MantisBT for real use. -> Elliott

    Retrieved from
    
"http://www.jlab.org/Hall-D/software/wiki/index.php/GlueX_Offline_Meeting%2C_June_15%2C_2010"

References

   29. 
http://www.jlab.org/Hall-D/software/wiki/index.php/Sim-Recon_Tagged_Releases
   30. 
http://www.jlab.org/Hall-D/software/wiki/index.php/How_HDGeant_defines_time-zero_for_physics_events
   32. 
http://www.jlab.org/Hall-D/software/wiki/index.php/GlueX_Offline_Meeting%2C_June_1%2C_2010#Minutes
   36. 
http://www.jlab.org/Hall-D/software/wiki/images/d/de/20100615_48cores.pdf