GlueX Offline Meeting, April 18, 2012

From GlueXWiki
Jump to: navigation, search

GlueX Offline Software Meeting
Wednesday, April 18, 2012
1:30 pm EDT
JLab: CEBAF Center, F326/327

Agenda

  1. Announcements
    1. EVIO version 4
    2. std::shared_ptr<>() and new C++ standard
  2. Review of minutes from the last meeting: all
  3. Reconstruction sub-group reports
    1. Calorimeters
    2. Tracking
      1. Tracking/FDC Segment code changed to read target z-position and length from XML geometry (before: z-position was hard-coded to 65.0 cm)
      2. CDC Candidate factory: no longer produces spurious matches between axial layers 1 and 5 when there are no matching hits in layer 3
    3. PID
  4. Offline Software Review Planning
  5. Action Item Review
  6. Review of recent repository activity: all

Communication Information

Video Conferencing

Slides

Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2012-2Q on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2012-2Q/ .

Minutes

Present:

  • CMU: Will Levine, Paul Mattione
  • IU: Ryan, Mitchell, Kei Moriya, Matt Shepherd
  • JLab: Mark Ito (chair), David Lawrence, Irina Semenova, Simon Taylor, Elliott Wolin

Announcements

  • EVIO version 4 is out: Elliott

This version has a unified interface to files, buffers, and sockets. The details of the implementation are hidden. Dictionaries have been added.

  • The new C++ standard and shared pointers: Elliott

A new version of the C++ standard has been released. It is not commonly included in OS distributions yet. The currently available compilers have special namespaces for auxiliary packages like boost and TRL. Much of that functionality is part of the new standard. We are not ready to switch yet, but this is coming down the pike. Elliott is especially interested in the use of shared pointers, which implement a form of garbage collection.

Review of minutes from the last meeting

We reviewed the minutes from the April 4th meeting.

  • Sascha Somov had an issue with the new rad-hard-free FCAL that is in the simulation. He was using and old version of the code with new calibration constants with non-optimal results. Mark created an older version of the calibration constants for his use.
  • Simon fixed a problem where not all of the energy deposited in the TOF was added into the hit energy. This was causing spurious inefficiency in the TOF.

Reconstruction sub-group reports

Calorimeters

  • FCAL Cluster Block Minimum

Will reminded us about discussion about the minimum number of blocks needed to form an FCAL cluster. Richard had suggested a change in the threshold for blocks next to the seed block and Will wants to explore the effect of that approach. We decided that in the mean time we would go ahead and implement a 1-block minimum as a default rather than 2. This will be changeable at run time with a JANA parameter. Matt also suggested that a loose time cut be added since Will has demonstrated good results with such cuts. +/- 5 ns was suggested. Will agreed to implement these changes.

  • Modeling BCAL Sampling Fluctuations

David reported on an ad hoc meeting held last week with Irina, Andrei Semenov, Elton Smith, and Eugene Chudakov to discuss the BCAL. One topic was now we smear the BCAL energy for photons. Although we simulated shower development by default, the geometry in simulation does not have the individual layers of lead and scintillator, so sampling fluctuations are not modeled. One approach is to simply add the finer grained geometry, but that may increase execution time unacceptably. The other is to introduce angle dependent smearing in addition to that generated by HDGeant. This requires keeping track of the kinematics of the original photon that initiated the shower, which is not done now. Both approaches will be studied.

Tracking

Simon gave the report.

  • The tracking/FDC Segment code changed to read target z-position and length from the XML geometry. Before z-position was hard-coded to 65.0 cm in many places. Matt recalled for us that the one of the reasons to put the target at a non-zero position was so that people would not hard code the target position. So much for that plan.
  • The CDC Candidate factory no longer produces spurious matches between axial layers 1 and 5 when there are no matching hits in layer 3.

PID

  • Paul is working on spiraling pions. Things are going well; the code is running. He is doing tests and dealing with problem cases.
  • David added DPiPlus and DPiMinus classes with associated factories to sim-recon. "These are convenience objects that place all reconstructed tracks for which the pi+(-) is the most probable into a single list."

Software Review preparations

David brought us up-to-date:

  • The committee chair and members have been named: Van Haus(chair), Marchand, McBride, Purske.
  • Another member may be named, perhaps from Babar or CLEO
  • No charge to the committee has been released yet.
  • Data acquisition has been added to the scope of the review.
  • Discussions between the organizers and the chair have taken place.
  • Chair want to make sure we understood how much work it takes to utilize the Grid. We need to have numbers prepared.
  • Outlines of slides have been done
    • Introduction: Curtis Meyer
    • Computing requirements: Mark
    • Software Infrastructure: David
  • There will be another meeting next week with Rolf.
  • The dates of the review remain June 7 and 8.
  • An outline of the document for hand-out to the committee has been drafted. This will serve as the basis for a section of the TDR.

Action Item and Recent Check-in's

We reviewed them.