GlueX Software Meeting, August 6, 2019
GlueX Software Meeting
Tuesday, August 6, 2019
3:00 pm EDT
JLab: CEBAF Center A110
BlueJeans: 968 592 007
- Review of minutes from the last Software Meeting (all)
- Event Generator Updates (Sean)
- Review of recent issues and pull requests:
- Review of recent discussion on the GlueX Software Help List (all)
- Action Item Review (all)
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/ .
- FSU: Sean Dobbs
- JLab: Mark Ito (chair), David Lawrence, Simon Taylor, Beni Zihlmann
The meeting was called to order at 3:00 pm
There is a recording of this meeting on the BlueJeans site. Use your JLab credentials to access it.
David gave us a status update.
- Test jobs submitted last Thursday did not run. They were canceled by NERSC staff for not-well-specified reasons.
- The jobs were re-submitted yesterday. Those too are at risk of being canceled. The problem is under investigation.
- Our jobs seem to be running into a problem with DVS, the NERSC system for compute nodes to access central disks.
- A few jobs did get through and started. They ended after running up against the 7.5 hour time limit. A small scale test of event processing rate showed no change from that previously measured in other launches.
- On the bright side, a transfer rate of 10 Gb/s was observed for several hours at a time.
- Cache disk outage on Aug. 8. Mark reminded us about the outage.
- Version set for reconstruction of Fall 2018 data at NERSC: recon-2018_08-ver02_1.xml.
- New update version set: version_4.7.0.xml. Includes new versions of AmpTools and RCDB.
- MCwrapper-bot issues. There was a delay in processing simulation jobs due to site configuration issues.
- New release of build_scripts: version 1.51. New packages hepmc, photos, and evtgen introduced with this release.
Review of minutes from the last Software Meeting
We went over the minutes from July 23.
- Flat Launches. Analysis launches to produce "flattened" analysis trees were discussed at the last Production and Analysis meeting. Since then Alex Austregesilo has changed the reaction submission page and Sean has changed the output data format to included unused neutral hypotheses. More work is needed before implementation of the new style of launches.
- The problem with multi-run reconstruction jobs remains an issue.
Event Generator Updates
Sean described recent work on simplifying generation of prompt particle decays in all of our event generators. See his slides for the details.
- Particle decays are handled by the EvtGen package. This package is used by many high-energy collaborations and benefits from years of development effort. Control of decays is configurable via simple text files.
- Decays are handled as a post-processor to any event generator that produces HDDM formatted output. Generated events are read in, particles decayed, and the resulting event written out in a new HDDM file. The result can be fed to HDGeant4. This mechanism works for many event generators without modification of the original code.
- The modification to build_scripts mentioned above in the announcements will build the packages needed to support this scheme.
- Implementation of the post-processor is now on a branch of the halld_sim repository. Sean has modified SCons on that branch to build it as an optional build product. He will be submitting a pull request soon.
Review of recent issues and pull requests
We did not do the review.
Review of recent discussion on the GlueX Software Help List
We skipped this review as well.
We adjourned at 3:58 pm.