<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=utf-8">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <p>Folks,</p>
    <p>Please find the minutes below and at
<a class="moz-txt-link-freetext" href="https://halldweb.jlab.org/wiki/index.php/GlueX_Offline_Meeting,_July_20,_2016#Minutes">https://halldweb.jlab.org/wiki/index.php/GlueX_Offline_Meeting,_July_20,_2016#Minutes</a>
      .</p>
    <p>--  Mark</p>
    <p>_____________________________</p>
    <p>
    </p>
    <div id="globalWrapper">
      <div id="column-content">
        <div id="content" class="mw-body" role="main">
          <h2 id="firstHeading" class="firstHeading" lang="en"><span
              dir="auto">GlueX Offline Meeting, July 20, 2016, </span><span
              class="mw-headline" id="Minutes">Minutes</span></h2>
          <div id="bodyContent" class="mw-body-content">
            <div id="mw-content-text" dir="ltr" class="mw-content-ltr"
              lang="en">
              <p>Present:
              </p>
              <ul>
                <li> <b>CMU</b>: Mike Staib</li>
                <li> <b>FIU</b>: Mahmoud Kamel</li>
                <li> <b>FSU</b>: Brad Cannon</li>
                <li> <b>JLab</b>: Alexander Austregesilo, Amber
                  Boehnlein, Mark Ito (chair), David Lawrence, Paul
                  Mattione, Nathan Sparks, Justin Stevens, Simon Taylor</li>
                <li> <b>NU</b>: Sean Dobbs</li>
                <li> <b>Regina</b>: Tegan Beattie</li>
                <li> <b>UConn</b>: Richard Jones</li>
              </ul>
              <p>You can <a rel="nofollow" class="external text"
                  href="https://bluejeans.com/s/a1dH/">view a recording
                  of this meeting</a> on the BlueJeans site.
              </p>
              <h3><span class="mw-headline" id="Announcements">Announcements</span></h3>
              <ol>
                <li> <b>dc3 retired</b>. Mark has officially put Data
                  Challenge 3 out of its misery. The <a
href="https://halldweb.jlab.org/wiki/index.php/GlueX-related_shared_accounts_on_the_JLab_CUE#Uses_of_the_GlueX_Project_accounts"
                    title="GlueX-related shared accounts on the JLab
                    CUE">gxproj2 account is free now</a>.</li>
                <li> <b>Analysis Launch</b>: Paul reminded us that
                  plugins for the analysis launch are <a rel="nofollow"
                    class="external text"
href="https://mailman.jlab.org/pipermail/halld-physics/2016-July/000705.html">due
                    this Friday at 5 pm</a>.
                  <ul>
                    <li> The topic sparked a discussion of how to manage
                      the ever growing number of plugins in sim-recon.
                      They have now significantly increase the build
                      time and build build disk footprint. Also many
                      perform important, central functions, some are
                      very close to private code. Some kind of
                      management scheme is called for. Note that this is
                      in some ways the kind of problem we want to have;
                      too much coding going on out there. Some ideas:
                      <ol>
                        <li> Move plugins to a separate plugins-only Git
                          repository. (Mark)</li>
                        <li> Reduce the number of plugins that are built
                          automatically. (Justin)</li>
                        <li> Have finer-grained build targets for SCons
                          so that parts of the build can be skipped
                          depending on the needs of the user. (Sean)</li>
                        <li> Break sim-recon into functionally distinct
                          (though interdependent) parts: simulation,
                          reconstruction, and analysis (Nathan)
                          <ul>
                            <li> Dependence on AmpTools might also need
                              some consideration in such a scheme
                              (Justin)</li>
                          </ul>
                        </li>
                      </ol>
                    </li>
                  </ul>
                </li>
              </ol>
              <h3><span class="mw-headline"
                  id="Review_of_minutes_from_the_last_meeting">Review of
                  minutes from the last meeting</span></h3>
              <p>We went over the <a
href="https://halldweb.jlab.org/wiki/index.php/GlueX_Offline_Meeting,_July_6,_2016#Minutes"
                  title="GlueX Offline Meeting, July 6, 2016">minutes
                  from the meeting on July 6</a>.
              </p>
              <ul>
                <li> After the Lustre upgrade, the performance of the
                  system has been a lot steadier. There have not been
                  any reported hangs of the ifarm since. Justin and
                  Sascha Somov have reported irregularities, but those
                  were cleaned up by IT without loss of data.</li>
                <li> sim-recon version 2.2.0 has been released. It fixes
                  a problem with the non-linear energy correction in the
                  FCAL and is being used in sim1.1.</li>
                <li> REST data from phase 2 of the data reconstruction
                  launches has been distributed to institutions outside
                  the Lab.</li>
                <li> Paul reported that the I/O contention reported by
                  Mike when reading HDDM-formatted data is not an issue
                  with the upcoming Analysis Launch. This is probably
                  because of the large number of plugins in the Launch
                  configuration.</li>
              </ul>
              <h4><span class="mw-headline" id="ROOT_6_Upgrade">ROOT 6
                  Upgrade</span></h4>
              <p>It looks like we are not ready to pull the trigger on
                this yet. Richard reported a problem building ROOT 6
                with GCC 4.9.2. Nathan will consult with him on tracking
                it down. There will be dislocations due to old macros
                and private code not working when the change is made,
                but it is something we will have to deal with sooner or
                later. Justin and Sean pointed out we need to make the
                switch well in advance of the Fall run if at all. We
                will revisit the issue in two weeks time.
              </p>
              <h3><span class="mw-headline"
                  id="Multi-threaded_REST_file_i.2Fo_with_parallel_hddm">Multi-threaded
                  REST file i/o with parallel hddm</span></h3>
              <p>Richard has implemented multi-threaded input and output
                in HDDM. It helps when the time spent in
                compression/de-compression is a significant fraction of
                the processing load. Implementation in JANA appears to
                require some significant effort and those involved have
                judged it not of the highest priority at present.
                Richard noted that in the case of pure Python
                implementations, the benefit will obtain.
              </p>
              <h3><span class="mw-headline" id="sim1.1">sim1.1</span></h3>
              <p>Sean reported that Mark has started and he is about 1/3
                of the way through after a couple of days. We are
                running version 2.2.0 of sim-recon. [Added in press: the
                jobs had been running with the wrong version. The
                version was corrected (set to 2.2.0) and the jobs
                restarted Wednesday evening.] Data is being simulated
                with runs matching those in the real data with 1% of the
                number of read-data events. In addition to REST files,
                monitoring histograms, and a fraction of the
                unreconstructed hddm-smeared data, trees from efficiency
                plugins are being kept.
              </p>
              <p>Sean looked at the histograms and they seem reasonable
                for the most part. There are some problems, however:
              </p>
              <ul>
                <li> BCAL timing is off. This has been seen for a while.</li>
                <li> <a rel="nofollow" class="external text"
href="https://halldweb.jlab.org/wiki/images/3/3c/Sim1.1_r10691_SC_Match.png">Start
                    Counter matching is terrible</a>, 60-75% depending
                  on how it is plotted. The timing may not be tuned in
                  the MC variation of CCDB.</li>
              </ul>
              <p>Alex A. asked why the code is not identical to that
                used in the reconstruction launches. Mark said that this
                issue had been discussed, but the reconstruction era
                code did not to the FCAL energy correction correctly for
                simulation, and the effort was not spent to cherry-pick
                the changes. Paul thought that that effort would have
                been wasted anyway; we will re-do reconstruction and
                simulation before too long anyway. We agreed that this
                is a reasonable attitude right now.
              </p>
              <p>Mark went through the <a rel="nofollow"
                  class="external text"
                  href="https://markito3.wordpress.com/2016/07/18/sim1-1-notes-2/">list
                  of plugins dropped</a> from sim1.1 for various
                reasons.
              </p>
              <p>For details of the conditions see the [<a
                  rel="nofollow" class="external free"
                  href="https://halldweb.jlab.org/gluex_simulations/sim1.1/"><a class="moz-txt-link-freetext" href="https://halldweb.jlab.org/gluex_simulations/sim1.1/">https://halldweb.jlab.org/gluex_simulations/sim1.1/</a></a>
                sim1.1 Static Website.
              </p>
              <h3><span class="mw-headline"
                  id="Changes_to_mcsmear_in_sim-recon">Changes to
                  mcsmear in sim-recon</span></h3>
              <p>Sean led us through <a rel="nofollow" class="external
                  text"
href="https://mailman.jlab.org/pipermail/halld-offline/2016-July/002419.html">his
                  email describing the changes</a>. Specifically he
                mentioned:
              </p>
              <ul>
                <li> TOF and Start Counter smearing changes</li>
                <li> Implementation of channel-by-channel efficiencies</li>
                <li> Use of the RCDB interface to fetch relevant
                  parameters directly.</li>
              </ul>
              <h3><span class="mw-headline" id="Thresholds:_CDC.2C_FDC">Thresholds:
                  CDC, FDC</span></h3>
              <p>Sean encouraged the detector groups to review the way
                threshold are applied in the Monte Carlo. There may be
                some further design needed now that we have real data to
                compare to. See the relevant sections on the <a
                  href="https://halldweb.jlab.org/wiki/index.php/Mcsmear_updates"
                  title="Mcsmear updates">mcsmear updates wiki page</a>.
              </p>
            </div>
            <div class="printfooter">
              Retrieved from "<a dir="ltr"
href="https://halldweb.jlab.org/wiki/index.php?title=GlueX_Offline_Meeting,_July_20,_2016&oldid=76490">https://halldweb.jlab.org/wiki/index.php?title=GlueX_Offline_Meeting,_July_20,_2016&oldid=76490</a>"</div>
          </div>
        </div>
      </div>
      <div id="footer" role="contentinfo">
        <ul id="f-list">
          <li id="lastmod"> This page was last modified on 23 July 2016,
            at 14:19.</li>
        </ul>
      </div>
    </div>
  </body>
</html>