Notes on 2013 Software Review
From GlueXWiki
- Round 2: 1st time 18 months ago
- Convened by Bob McKeown
- Committee:
- Torre Wenaus (BNL/ATLAS) - Committee Chair
- Elizabeth Sexton (FNAL/CMS)
- Sergei Gerassimov (CERN)
- Martin Purschke (BNL/PHENIX)
- David Nathan Brown (LBL/BaBar)
- talks:
- Curtis, overview
- Mark, computing plan, data challenges
- David, improvements in reconstruction
- Justin, analysis workshop: event generators, analysis tools, event classification (BDT), amplitude analysis with AmpTools
- written material:
- GlueX Computing document, updated from last time
- Appendices:
- data management plan, Lab and Hall D
- draft analysis plan
- workflow system requirements
- Committee summary:
- cited excellent progress for all Halls during the past 18 months and thought we are on track for data taking
- For the 12 GeV program as a whole
- they noted the use of parallel processing schemes,
- encouraged more use of static and dynamic code profiling.
- They asked that we consider doing code review for crucial new developments.
- They also liked that software workshops that have been held, noting that recording of the sessions is valuable for reaching a larger audience.
- For Hall D
- they noted the data challenges we have performed,
- the fact that the data from DC I was used to support the PAC 40 proposal.
- They also mentioned the inclusion of QA and data challenge management in our software manpower plan
- applauded the appointment of a Calibration Coordinator.
- suggestion to tag events online as good candidates for calibration input
- Report:
- An interim report will be made available to Lab management soon
- the Committee hopes to issue its final report before the end of the year.
- Personal Comments:
- we were the only data challenges
- probably the most comprehensive written material
- no push back on using JLab-specific tools for JLab farm workflow
- suggestions on tools: no SRM for data transfer, no LFC file catalog
- updating tools: can get up-to-date advice, take advantages of recent developments
- lot of time spent talking about computer center issues
- overall, easy on us
