GlueX Offline Meeting, January 12, 2011

From GlueXWiki
Revision as of 11:53, 13 January 2011 by Marki (Talk | contribs) (complete draft)

Jump to: navigation, search

GlueX Offline Software Meeting
Wednesday, January 12, 2011
1:30 pm EST
JLab: CEBAF Center F326/7

Agenda

  1. Announcements
  2. Review of minutes from the last meeting: all
  3. Impressions from the Parallelism Workshop: David
  4. IT Readiness for 12 GeV: Graham Heyes
  5. Changes to TOF code: Beni
  6. Tracking Issues: Kei
  7. Action Item Review: all
    1. Resolved issues
    2. Outstanding issues
    • To view these links, login first. Anonymous login is available.
  8. Review of recent repository activity: all

Communication Information

Telephone

To connect by telephone:

  1. dial:
  2. enter access code followed by the # sign: 3421244#

Video Conferencing

Slides

Talks can be deposited in the directory /group/halld/www/halldweb1/html/talks/2011-1Q on the JLab CUE [1]. This directory is accessible from the web at https://halldweb1.jlab.org/talks/2011-1Q/ .

Minutes

Present:

  • FSU: Nathan Sparks
  • IU: Jake Bennett, Ryan Mitchell, Kei Moriya, Matt Shepherd
  • JLab: Hovanes Egiyan, Graham Heyes, Mark Ito (chair), David Lawrence, Yi Qiang, Dmitry Romanov, Simon Taylor, Beni Zihlmann

Announcements

Mark encouraged people to start thinking about what talks they want to give at the upcoming Collaboration Meeting in the Offline session.

Parallelism Workshop

David gave some impressions from last week's Workshop on Parallelism in Experimental Nuclear Physics. Fifty five people registered and a like number attended. There were lots of interesting and informative talks, featured among them talks from folks outside the JLab community. See David's email for his summary for the participants and a link to the virtual machine that has the hands-on examples loaded on it.

In discussion Matt asked if JLab has licenses for Purify. Graham said that it did; he would look into who owns it now.

Review of IT Readiness for the 12 GeV Era

Graham described the upcoming review of all computing plans for 12 GeV era. The review will be comprehensive in that it will include accelerator controls, data acquisition support software, the Computer and Network Infrastructure Department (CNI) and the Management Information Systems Department (MIS). Some of his points:

  • want to determine if we can handle data coming from experiments
  • need an outline of remarks by all speakers by end of month
  • the review will be around the 2nd week of March
  • Graham will send around a spreadsheet, asking for updates to computing requirements from various groups
  • scheduled speakers: Bob Michaels from Hall A, Dennis Weygand from Hall B, Mark Ito from Hall D, a representative from Hall C--TBA, and Graham with an overview
  • current computing requirements:
    • Hall D: compute intensive
    • Hall B: storage intensive
  • storage more of a worry
  • this will be an internal (to JLab) review
  • a pre-cursor to external review in a year or so

Changes to TOF Code

Beni described recent changes he has made to the HDDM data model to accommodate information from the Monte Carlo about "truth" information related to time-of-flight hits. In some cases, new elements were introduced to be able to keep both generated and detected information, and some elements were expanded to store information about the particle producing the hits.

A lot of discussion followed on the general idea of keeping generated quantities ("truth") and detected quantities ("hits") separate in our analysis. Some of Beni's changes seemed to violate this separation. Beni, Mark, and David agreed to meet after the meeting to discuss alternate approaches that preserve the information desired, but make the above mentioned separation more explicit.

Tracking Issues

Kei presented a list of bugs/anomalies that he has come across in running our reconstruction code. See his talk for details. He mentioned four items:

0. event processing hangs every ~1000 events for his event topology
1. pi+ events reconstructed as proton show “spike” in momentum at ~0.5 GeV/c
2. some events have direction of momentum “reversed”
3. values of variables are set to non-physical values when they cannot be defined for a given track (leads to side-effects if not checked for explicity)

Work continues to try to resolve these mysteries.

Side note: Mark encouraged folks to write to halld-offline@jlab.org when they encounter problems like this so that (a) we all are aware of the problem (b) others that encounter similar problems can benefit from the conversation.

Action Item Review

We looked over recently resolved items and those outstanding. David has added some items based on feedback from the workshop. Mark got David to volunteer to update the Build Make System (BMS) document to bring it up-to-date.

New Action Items

  1. Look into the Purify license. -> Graham
  2. Send email to the list on recent problems in the offline code. -> Kei
  3. Prepare revised computing requirements for IT review. -> Mark
  4. Revise BMS document. -> David

Footnotes

  1. You have to be a member of the "halld" Unix group to do this.