GlueX Offline Meeting, November 14, 2012
GlueX Offline Software Meeting
Wednesday, November 14, 2012
1:30 pm EDT
JLab: CEBAF Center, F326/327
- 1 Agenda
- 2 Communication Information
- 3 Minutes
- 3.1 Announcements
- 3.2 Detector naming scheme
- 3.3 More on raw data, occupancies and anomalies
- 3.4 Review of minutes from the last meeting, geometry changes
- 3.5 Report from the last Data Challenge Meeting
- 3.6 Tracking
- 3.7 Geometric consistency
- 3.8 Resources
- Review of minutes from the last meeting: all
- Detector naming scheme (from Fernando): Elliott
- More on raw data, occupancies and anomalies: Elliott
- Report from the last Data Challenge Meeting
- Reconstruction sub-group reports
- Geometric consistency(slides): David
- Resources(slides): David
- Detector Nomenclature: Elliott
- Action Item Review
- Review of recent repository activity: all
Talks can be deposited in the directory
/group/halld/www/halldweb/html/talks/2012-4Q on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2012-4Q/ .
- CMU: Will Levine, Paul Mattione, Curtis Meyer
- JLab: Mark Ito (chair), David Lawrence, Dmitry Romanov, Sascha Somov, Simon Taylor
Mark called our attention to a recent announcement from Dmitry and him on the latest release of the calibration database. The code is built on three platforms at JLab and there are instructions on how to use it with Jana in the file ccdb_0.05/janaccdb/README.txt.
Major changes on JLab batch farm
Mark went over the recent announcement from Scientific Computing at JLab on changes to the batch farm going in today and possibly tomorrow. There are many fundamental differences being rolled out in Auger and Jasmine so users should be on the lookout for anomalous behavior and report any seen. You will have to have a new scientific computing certificate to use the farm.
Scientific Software Support Committee
Mark told us about a new committee that will oversee support for scientific software packages at the Lab. Chip Watson proposed the existence of the committee and the first meeting was held last week. The committee will be joint between Physics and IT. Graham Heyes will act as chair. The initial initiative will be to integrate ROOT, CERNLIB, Geant4, etc. into the normal support structure of IT Division, mainly helpdesk service and documentation. The actual support of the packages will remain with the current responsible individuals for now. though that may change going forward. In the future the committee may help decide which packages should be officially support by IT.
Detector naming scheme
Elliott called our attention to the detector abbreviation scheme Fernando Barbosa has urged us to follow. He noted that there are several instances where the naming scheme in the code does not match that of Fernando's. We thought that consistency would be nice, but questioned whether this was an important front-burner issue for this group. We decided to back-burner it and perhaps discuss it again in the future.
More on raw data, occupancies and anomalies
Elliott described recent progress on online monitoring using the new EVIO format data. He outlined plans and showed histograms from the prototype monitoring plug-ins he has written. He solicits feed back and contributions from the detector groups to develop their respective packages further.
Review of minutes from the last meeting, geometry changes
We went over the minutes of the October 31st meeting.
David recently made some changes to the BCAL geometry to fix volume overlaps reported recently by Simon and by Richard some months ago. Simon also checked in a change to the FDC geometry so that it matches the inner radius of the BCAL. David did a radiation length scan and saw an increase in the amount of material inside the BCAL caused by the cable material being "revealed" when overlap was removed.
Report from the last Data Challenge Meeting
We looked over the minutes from the November 5th meeting.
Mark showed the web page that calls out conditions for the upcoming data challenge. He has done small scale tests with the tagged version of sim-recon specified there. Paul has started testing this version at CMU.
Simon has just checked in a modification to the CDC code to allow the geometry to be read in from the HDDS files rather than using hard-wired values in the code.
David described a recent feature that allows one to build all of the geometry routines from the HDDS package on the fly when building HDGeant. In this way, one can script the building of code with multiple geometries. This is especially useful when exploring the effect of a series of changes in a few parameters of interest. To further reduce the possibility of a mismatch between geometry and code, he stores a MDA5 sum in HDDM-s that "fingerprints" the geometry configuration that was used.
David described a new feature of Jana that supports the use of resources. In our case, these are files that are too large to be conveniently stored in the calibration database, like the magnetic field maps. The scheme allows the database to specify only URL of the file and not store the data itself. A local cache is maintained so that subsequent requests need not involve a network transfer. Mark and Dmitry urged David to support a use case where only a relative path and file name are returned from the calibration database. The web-site information and the absolute path of the root directory of the local cache then are controlled globally for a given job, outside of the database entry (the latter is already implemented, but the form is not). In this case the relative path and file name alone specify the file identity, both on the server and in the local cache. This eliminates the need for a local file that catalogs the state of the local cache. David agreed to make this change.