GlueX Offline Meeting, April 13, 2016

From GlueXWiki
Jump to: navigation, search

GlueX Offline Software Meeting
Wednesday, April 13, 2016
1:30 pm EDT
JLab: CEBAF Center L207

Agenda

  1. Announcements
    1. RCDB News
    2. Collaboration Meeting May 9-11, 2016
    3. Other announcements?
  2. Improved ROOT JANA Locks (Paul, David):
    1. Plugin-specific, action specific, factory specific
    2. New JANA release: 0.7.5
    3. Pull request #331
    4. Upgrade your JANA
    5. sim-recon version 1.11.0 released
    6. Locking in JANA
    7. New Scaling Study (Paul)
  3. Review of minutes from March 30 (all)
  4. Calibration Challenge/Processing (Sean)
  5. sim-recon scripts
  6. FADC Emulation Partial Re-Write (Richard)
  7. SWIF for GlueX Users
  8. Review of recent pull requests (all)
  9. Action Item Review

Communication Information

Remote Connection

Slides

Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2016 on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2016/ .

Minutes

There is a recording of this meeting on the BlueJeans site.

Present:

  • CMU: Naomi Jarvis, Curtis Meyer, Mike Staib
  • FIU: Mahmoud Kamel
  • JLab: Mark Ito (chair), Paul Mattione, Sandy Philpott, Dmitry Romanov, Nathan Sparks, Justin Stevens, Simon Taylor
  • NU: Sean Dobbs
Software Meeting

Announcements

  1. RCDB News
  2. Collaboration Meeting, May 9-11, 2016. The Collaboration Meeting Agenda gives the offline almost two hours.
    • Paul will give a talk on monitoring.
    • Sean will give a talk on calibration.
    • We will collectively think about other talks.

Improved JANA Locks

Paul explained the situation.

  1. The locks that have been modified are specific to different "areas" of the code; there is now differentiation between a special purpose locks and the global thread lock. The areas are:
    1. plugins
    2. actions
    3. factories
  2. The new scheme requires use of a new JANA release: 0.7.5.
  3. The crucial change came in with Pull request #331.
  4. Paul sent an email on Monday warning us about the need for use of the new JANA release.
  5. Mark released a new version of sim-recon, version 1.11.0, that incorporates these changes.
  6. Paul put together a wiki page giving best practices for using locks with JANA, inspired by the code he discovered during his review of the locks in our code.
  7. Paul also showed some slides contrasting the new lock discipline versus the lack thereof showing much improved multi-thread scaling.

Review of minutes from March 30

We went over the minutes from March 30.

  • Future Trends in Nuclear Physics Computing Workshop. Amber and Mark had a conference call with Richard Jones, our virtual organization representative, to plan out the next step in enhancing JLab involvement with the Open Science Grid.
  • Access to Fairshare Information from JLab Farm. We asked Sandy about getting the current fairshare statistics posted somewhere. She told us that it will be there on the new webpages that SciComp is working on. Mark suggested that at some point someone from SciComp should give presentation to this group about the new site.

Calibration Challenge/Processing

Sean presented a slide bringing us up-to-date with calibration activity.

  • He described the parameters of the three launches started so far. One set of statistics not on the slide:
Launch Number Number of Farm Jobs
1 6,000
2 3,000
3 3,500
  • The CCDB improvements he has discussed with Dmitry address ease of handling of calibration constants and possible calibration visualization ideas.
  • The benchmark jobs are to explore performance as a function of the number of threads. Sean is hoping that Paul's recent improvements will help the situation here as well.

sim-recon scripts

Paul has proposed moving the scripts that generate analysis plugin templates from the scripts directory in the Subversion repository to a new directory of the sim-recon repository (Git). That way these scripts will be versioned along with releases of sim-recon. That, in turn, would disambiguate which version of the scripts should be used with which version of sim-recon.

We all agreed that this is a good idea, but Nathan wanted to think about whether there might be a more appropriate location than that proposed by Paul. We agreed to let him make an alternate proposal before we make the move.

FADC Emulation Partial Re-Write

We looked at Richard's recent pull request. The situation seems to be well in hand.

SWIF for GlueX Users

Mahmoud reminded that Paul's guide to SWIF for GlueX Analysis Users had been discussed previously. He is using the system himself.