GlueX Offline Meeting, June 1, 2011
GlueX Offline Software Meeting
Wednesday, June 1, 2011
1:30 pm EDT
JLab: CEBAF Center A110
- 1 Agenda
- 2 Communication Information
- 3 Minutes
- Review of minutes from the last meeting: all
- DFCALPhoton vs. DFCALShower: Mihajlo
- b1pi fix: David, Paul
- Kinematic fitting in GlueX: Kei
- Re-cap of review, IT in the 12 GeV Era: Mark
- Grid Make System: Richard
- Action Item Review: all
- Review of recent repository activity: all
To connect by telephone:
- US and Canada: (866)740-1260 (toll free)
- International: (303)248-0285 (toll call) or look up toll-free number at http://www.readytalk.com/intl
- enter access code followed by the # sign: 3421244#
- ESNet: 8542553
- Meeting URL: http://evo.caltech.edu/evoNext/koala.jnlp?meeting=eueneevIv9aiavavIu
- Phone Bridge ID: 8 6750
Talks can be deposited in the directory
/group/halld/www/halldweb/html/talks/2011-2Q on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2011-2Q/ . You have to be a member of the "halld" Unix group to do this.
- UConn: Alex Barnes, Richard Jones
- IU: Mihajlo Kornicer, Kei Moriya
- JLab: Hovanes Egiyan, Mark Ito (chair), David Lawrence, Paul Mattione, Yi Qiang, Sascha Somov, Ivan Tolstukhin, Elliott Wolin, Beni Zihlmann
At Elliott's urging, Ivan announced that Dmitry Romanov will return to JLab mid-June, and plans to stay for six months. Ivan himself will be leaving mid-July.
Review of Previous Meeting's Minutes
We went over the minutes from the May 18th meeting.
- Beni will give a talk on the new primary/secondary particle identification scheme in HDGeant output at the next meeting.
- We wondered again if there was any progress with a figure/plot library.
DFCALPhoton vs. DFCALShower
Mihajlo and David explained the situation.
- There is an inconsistency in the naming convention between the FCAL and the BCAL.
- Mihajlo will work on bringing them into registration.
- The hierarchy is
- detector hit
- A shower does not become a particle until associated with a vertex. The photon has a full three-momentum. Before that the shower only contains local coordinate information.
- Mihajlo proposed that the algorithm to merge showers should be applied after matching showers with charged tracks and not before matching. David and Mihajlo will discuss this offline and report to the group.
propose to make this change.
At some point in the past few months the reconstruction efficiency for the weekly automatic b1pi analysis started to show only 5% reconstruction efficiency when it used to be around 50%. The problem was found and fixed. Paul showed a wiki page illustrating the new situation. Shown there are various invariant mass plots, and identified particle multiplicity distributions.
David explained that there were two problems:
- The vertex factory, where both charged and neutral particles are grouped into vertices had a hard-wired cut at 3 σ in both time and z. Several vertices were being found. The cut was widened considerably, so that only one vertex is almost guaranteed. The cut can now be set with a configuration parameter. This will have to be worked on further.
- In the DVertex class there is a member class that inherits from DKinematicData that has pointers to cluster objects that were not being copied to the DKinematicData member data. This is being done now.
Kinematic fitting in GlueX
Kei gave a review of the imponderables he has encountered trying to complete an exercise in kinematic fitting with our simulation and reconstruction code. See his slides for details. The group suggested that the issues that are causing trouble might be more easily addressed at a more fundamental level in the reconstruction code.
Re-cap of Review, IT in the 12 GeV Era
Mark went over the close-out slides from the review conducted on May 20. For copies of the talks and the charge to the committee see the review web page. A couple of points that come up in discussion:
- Mark claimed that if Hall D's requirements seemed less certain than those of B, it was only because the uncertainty was exposed, not that Hall B's requirements were more certain.
- Richard remarked that we should use the feedback from the committee constructively and be well-prepared when the external review comes in a year or so.
Grid Make System
Richard describe a system he is developing to
- automatically install and build the GlueX software suite on an arbitrary Unix-based system and
- use the build to create user-specified output, the result of either simulation, reconstruction, and/or analysis.
It uses a make-style paradigm where targets are specified, those targets have pre-requisite targets, and any pre-requisite that does not exist gets made. The system is driven by a perl script. Recipes for creating targets are kept in separate files, target-by-target. The configuration of the system is done via XML. Version choices for the various software components can be enforced. The code is now in alpha release. See his slides for details.
New Action Items
- Track progress on the figure/plot library for GlueX
- Decide on cluster merging after or before charged track matching.
- Do further development on the gridmake system.