GlueX Offline Meeting, June 1, 2010
GlueX Offline Meeting
Tuesday, June 1, 2010
2:00 pm EDT
JLab: CEBAF Center F326/7
To connect by telephone:
- 800-377-8846 : US
- 888-276-7715 : Canada
- 302-709-8424 : International
- enter participant code: 77438230# (remember the "#")
- ESNet: 8542553
- Meeting URL: http://evo.caltech.edu/evoGate/koala.jnlp?meeting=eleBevvlvDatasaeIl
- Phone Bridge ID: 56083
Talks can be deposited in the directory
/group/halld/www/halldweb/html/talks/2010-06 on the JLab CUE (you have to be a member of the "halld" Unix group). This directory is accessible from the web at https://halldweb.jlab.org/talks/2010-06/ .
- Can we make binaries-in-path a requirement? Yes.
- Modifications to coding conventions done, added warnings and random stuff in mcsmear only
- coding conventions page needs revision, volunteers?
- Automate links from CentOS5 to RHEL5-64-bit, done, script in scripts/build_scripts
- b1pi analysis being run in cron job
- HOWTO do a kinematic fit for etapi0p events: Blake
- BMS document linked from software wiki page
- HOWTO use a pre-built release: Mark
- Review of minutes from the last meeting: all
- News from IU
- GlueX Software Workshop: David
- Bug/issue tracking software: Elliott
- Effect of a more uniform field; raised by Eugene at GlueX meeting
- Action Item Review: all
JLab: Craig Bookwalter, Mark Ito (chair), David Lawrence, Simon Taylor, Elliott Wolin, Beni Zihlmann
- Can we make binaries-in-path a requirement? Yes. See Mark's email message. This issue was raised by the desire to remove a hard-wired path reference in the makefile for sim-recon/src/libraries/HDDM.
- Modifications to coding conventions done. Mark added the policy on warning-free code the policy on random stuff being restricted to the "mcsmear" step in Monte-Carlo generation. The wiki page needs to be cleaned up. No volunteers where found.
- Automate links from CentOS5 to RHEL5-64-bit. Mark wrote a script to do this semi-automatically.
- analysis being run in a cron job. Mark has David's script running weekly. It runs on all three platforms presently.
- David mentioned that the next step would be to create histograms. Right now a root tree is produced but the only information available is that the reconstruction ran to completion. He has a script that will produce histograms and mail out the results. He and Mark will create an example that runs on the cron job output.
- HOWTO do a kinematic fit for etapi0p events. We collectively took a glance at Blake Leverington's recent HOWTO. See his email for a synopsis.
- BMS document linked from software wiki page. David pointed Mark to the appropriate GlueX Doc and it is now linked from the Offline Software wiki page. There are parts of this that need revision. No volunteers were found.
- HOWTO use a pre-built release. This HOWTO was announced last week on the email list. It describes our "standard procedure" for using a release of sim-recon that someone else has built.
- Release method for DAQ Group software is broken. Elliott told us that there are aspects of the current release method that cannot possibly work. This issue will be raised at the online meeting at some point.
Review of minutes from the last meeting
We reviewed the minutes from the May 18 meeting.
- ROOTSpy. The student from CNU that was slotted to work on ROOTSpy has reported for work, although Mark opined that this was an announcement more appropriate for the Online Meeting.
- Setting up the GlueX software environment. We reached a consensus that Mark should decide on what to recommend to the Collaboration as a best practice.
- Progress toward constructing our own matrix classes. Since the last meeting Simon has tried his new matrix classes in the Kalman fitter. He sees very nice improvement in execution times.
- Hardware/Connection tracking database. Mark suggested that this issue should probably be taken up at the Online Meetings.
News from IU
As reported by Matt Shepherd:
- Dan Bennett is making nice progress on the BCAL work. The first goal is move the "MCResponse" to the mcsmear application and also handle summing electronics channels that is presently in the design. That effort is somewhat complicated since it breaks the nice 1 to 1 or 1 to 2 (ends) mapping from GEANT to reconstruction. We are handling this by renaming all objections in the data model: sipmCell, sipmLayer, sipmSector, and fadcCell, fadcLayer, fadcSector. This will undoubtedly take some refining, I'm sure you'll be hearing updates from us in future meetings.
- On the amplitude analysis side of things, Matt Shepherd has a set of amplitudes from Adam Szczepaniak to do mock analysis on pi_1 -> rho pi channel. Matt hopes to have an example in place by the end of the week.
GlueX Software Workshop
At the last meeting, the idea of a GlueX Software workshop was discussed. Since then the CLAS12 Workshop was held; there were about 45 participants on the first day, about 30 attended the tutorials on the second day. Also David called a meeting of the committee that was formed, and discussions of how the workshop might be organized has started with other parties, notably Eugene Chudakov, Ed Brash, and Ole Hansen. An expanded committee will meet this week. There are a lot of ideas being discussed at present; the meeting should sort some of them out.
Bug/issue tracking software
Elliott outlined our options:
Mantis has been installed on halldnew.jlab.org, but is not currently visible off-site. We are waiting for Marty Wise of CNI to implement LDAP authentication before they will allow access from outside the Lab. Sherman White of CNI has installed a Drupal instance for us to test, but no one has messed with it yet. This is also only available inside the firewall.
Right now it looks like Mantis is the leading candidate. It has a lot features, many more than we are actually looking for, but still has an intuitive user interface. So that leaves room for growth, while not defeating our purposes.
Effect of a more uniform field
David and Simon are doing studies aimed at developing requirements should we decide to re-design the solenoid magnet.
- Simon is looking at using a smaller step size in the swimming, although he thinks that the effect of material dictates a smaller step size that dictated by the non-uniformity in magnetic field.
- David will look at the efficiency of track finding with a more uniform field.
Mark reported that he spoke with Eugene after the last GlueX Meeting. Mark thinks that Eugene wants to see whether a non-swimming approach to track fitting is useful, at least in some contexts, if the field was much more uniform. This would greatly increase reconstruction speed.
Action item review
We went over the list. People were reminded of things that they signed up to do.
New Action Items
- Clean up coding conventions wiki page.
- Add analysis of b1pi output to cron jobs -> Mark and David
- Decide on best practice for environment set-up -> Mark
- Find a volunteer to revise the BMS document.
- Do studies of a more uniform magnetic field.
Recorded by Mark Ito, 2 June 2010