<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=UTF-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <p>Folks,</p>
    <p>Find the minutes <a moz-do-not-send="true"
href="https://halldweb.jlab.org/wiki/index.php/GlueX_Offline_Software_Meeting,_October_16,_2018#Minutes">here</a>
      and below.</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">Minutes, GlueX Offline Software Meeting,
              October 16, 2018</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> Naomi Jarvis</li>
                <li> <b> FIU: </b> Mahmoud Kamel</li>
                <li> <b> FSU: </b> Sean Dobbs</li>
                <li> <b> IU: </b> Ahmed Foda</li>
                <li> <b> JLab: </b> Alex Austregesilo, Thomas Britton,
                  Mark Ito (chair), David Lawrence, Simon Taylor, Beni
                  Zihlmann</li>
                <li> <b> W&M: </b> Justin Stevens</li>
              </ul>
              <p>There is a <a rel="nofollow" class="external text"
                  href="https://bluejeans.com/s/KkfFM/">recording of
                  this meeting</a> on the BlueJeans site. Use your JLab
                credentials to access it.
              </p>
              <h3><span class="mw-headline" id="Announcements">Announcements</span></h3>
              <ol>
                <li> New version of MCwrapper: <a rel="nofollow"
                    class="external text"
href="https://github.com/JeffersonLab/gluex_MCwrapper/releases/tag/v2.0.2">version
                    2.0.2</a>. The bot is now in beta!</li>
                <li> New version of build_scripts: <a rel="nofollow"
                    class="external text"
href="https://mailman.jlab.org/pipermail/halld-offline/2018-October/003394.html">version
                    1.4.2</a>. A two-stage build process is supported.</li>
                <li> <a rel="nofollow" class="external text"
href="https://mailman.jlab.org/pipermail/halld-offline/2018-October/003383.html">New
                    version of halld_recon: recon-ver03.2</a> This
                  version is being used in the monitoring launch at
                  NERSC.</li>
                <li> More new versions: <a rel="nofollow"
                    class="external text"
href="https://mailman.jlab.org/pipermail/halld-offline/2018-October/003385.html">version_3.7_jlab.xml
                    feat. halld-recon 3.2.0, halld-sim 3.5.0,
                    gluex_root_analysis 0.5</a> A periodic-code-update
                  version set.</li>
              </ol>
              <h3><span class="mw-headline"
                  id="Review_of_minutes_from_the_October_2_meeting">Review
                  of minutes from the October 2 meeting</span></h3>
              <p>We went over <a
href="https://halldweb.jlab.org/wiki/index.php/GlueX_Offline_Software_Meeting,_October_2,_2018#Minutes"
                  title="GlueX Offline Software Meeting, October 2,
                  2018">the minutes</a>.
              </p>
              <p>We spent some time discussing the issue on the
                kinematic fitter reported by Hao Li and Mike McCracken.
                See the <a rel="nofollow" class="external text"
href="https://groups.google.com/forum/#!topic/gluex-software/lBh4R6B0j68">thread
                  on the software help list</a>. Beni pointed out that
                this is an issue that this working group should take on.
              </p>
              <p>[Added in press: Sean created a <a rel="nofollow"
                  class="external text"
                  href="https://github.com/JeffersonLab/halld_recon/issues/39">GitHub
                  issue to track progress on this problem</a>. He
                assigned the issue to himself and Thomas.]
              </p>
              <h3><span class="mw-headline" id="Computing_Review">Computing
                  Review</span></h3>
              <p>David and Mark reported that they met with Curtis two
                weeks ago to start planning for the review. Curtis has
                put together <a
href="https://halldweb.jlab.org/wiki/index.php/Software_and_Computing_Review_5"
                  title="Software and Computing Review 5">a wiki page to
                  collect materials</a>. They started a list of topics
                to address in the short time allowed, most importantly
                an updated estimate of future computing resource needs.
                They also want to highlight recent use of off-site
                computing resources.
              </p>
              <h3><span class="mw-headline"
                  id="Getting_to_the_ROOT_of_things...">Getting to the
                  ROOT of things...</span></h3>
              <p>Mark reviewed the <a rel="nofollow" class="external
                  text"
href="https://mailman.jlab.org/pipermail/halld-offline/2018-October/003396.html">recent
                  email from Graham Heyes</a> on opening a communication
                channel between local ROOT users and the ROOT
                development team. Bob Michaels is organizing a meeting
                with Alex Naumann of CERN. The meeting will be announced
                when plans are set. Interested parties are welcome.
              </p>
              <h3><span class="mw-headline"
                  id="Hardware_needs.2C_feedback_to_Chip">Hardware
                  needs, feedback to Chip</span></h3>
              <p>We reviewed <a rel="nofollow" class="external text"
href="https://halldweb.jlab.org/talks_secure/2018/SciComp%20ENP%20Co-Design%20FY2019.pptx.pdf">Chip's
                  presentation (user=writer...(contact me for more
                  info))</a> on options we have for spending on
                computing resources. Several of us (Mark, David, Alex,
                Sean, Thomas, Curtis, and Richard) met with him about
                two weeks ago where he solicited input in what our needs
                are vis-a-vis FY19 equipment purchases. Discussion
                points:
              </p>
              <ul>
                <li> Alex told us that the 2017-01 data produced 11-0 TB
                  of REST files. The associated root trees are about the
                  same size.</li>
                <li> The 2018-01 data should be about three times the
                  size of 2017-01.</li>
                <li> We have been running into problems with limited
                  Lustre-based disk space for the past year or so.</li>
                <li> David told us the the launches at NERSC would
                  benefit from dedicated space to stage the raw data
                  files so that that effort does not compete with
                  others.</li>
              </ul>
              <p>In the end we settled on rough proportions of where we
                would like our share of resources to go (on a dollar
                basis):
              </p>
              <ul>
                <li> 50% Lustre-based disk space (about a petabyte of
                  space)</li>
                <li> 40% Computing nodes (16 40-core nodes (80
                  hyper-threads) or 5.6 million core hours per year)</li>
                <li> 10% SSD disk space (about 25 TB on top of existing
                  25 TB dedicated to raw data input staging, could be
                  much cheaper (i. e. more space) if other Halls want a
                  like amount)</li>
              </ul>
              <p>We would want to defer the purchase of the computing
                nodes under the assumption that a later purchase might
                get us more computing per dollar spent.
              </p>
              <p>The need for SSD disk space is not certain, but several
                applications might potentially benefit. This small
                fraction would serve to gain experience to see if more
                high-speed disk would help us.
              </p>
              <p>Mark will talk to Hall B about what they are planning.
              </p>
              <h3><span class="mw-headline"
                  id="Review_of_Offline_Work_Packages">Review of Offline
                  Work Packages</span></h3>
              <p>We went over the <a rel="nofollow" class="external
                  text"
href="https://halldweb.jlab.org/wiki-private/index.php/2018_Work_Packages">list
                  of Analysis Software Work Packages</a>. Mark did a
                pass at marking up the list and Sean commented on the
                mark-ups on <a rel="nofollow" class="external text"
href="https://halldweb.jlab.org/wiki-private/index.php/Talk:2018_Work_Packages">the
                  corresponding "discussion" page</a>.
              </p>
              <p>If people have ideas about other work packages, please
                add them to the list.
              </p>
              <p>Mark agreed to break the list into two categories
                (Analysis and General) and fill in names of those he
                knows will volunteer to supervise packages.
              </p>
              <h3><span class="mw-headline" id="Meta_topics">Meta topics</span></h3>
              <p>At the last meeting several topics were raised.
              </p>
              <h4><span class="mw-headline"
                  id="Encouraging_wider_participation_in_these_meetings">Encouraging
                  wider participation in these meetings</span></h4>
              <ul>
                <li> The work packages may be a way to get new
                  collaboration members to attend once they get
                  volunteered to do the work.</li>
                <li> Some of the expert-level discussion might not be of
                  general interest.</li>
                <li> Many of the topics discussed at other working group
                  meetings, especially the Analysis Working Group, might
                  be more appropriately hosted at this meeting. Those
                  topics should be identified by people who attend these
                  other meetings.</li>
              </ul>
              <h4><span class="mw-headline" id="Rename_this_meeting">Rename
                  this meeting</span></h4>
              <p>One of us (a.k.a. Naomi) argued that "Offline" in the
                title did not have an auspicious connotation, e. g.,
                "off the main line" or "off topic", i. e., "irrelevant".
                We formed a consensus around "Hall D Software" rather
                than "Offline Software". We are reprinting the business
                cards now.
              </p>
              <h4><span class="mw-headline" id="Tutorials_and_Workfests">Tutorials
                  and Workfests</span></h4>
              <p>Past gatherings have proved useful. There was a lot of
                discussion, but we arrived at
              </p>
              <ol>
                <li> having occasional workfests focused on specific
                  work packages limited to those interested the specific
                  topic.</li>
                <li> once a year, on the day before the Spring
                  Collaboration Meeting, having a half-day software
                  tutorial like we did last Spring.</li>
              </ol>
              <h3><span class="mw-headline"
                  id="Review_of_recent_pull_requests">Review of recent
                  pull requests</span></h3>
              <p>David explained <a rel="nofollow" class="external
                  text"
                  href="https://github.com/JeffersonLab/halld_recon/pull/38">his
                  recent pull request (#38)</a>. There is now an option
                whereby DANA applications can create a local copy of the
                CCDB SQLite file indicated in JANA_CALIB_URL. This will
                help with slow start up, due to multiple processes
                hitting a single file, on the monitoring farm and other
                similar applications.
              </p>
            </div>
            <div class="printfooter">
              Retrieved from "<a dir="ltr"
href="https://halldweb.jlab.org/wiki/index.php?title=GlueX_Offline_Software_Meeting,_October_16,_2018&oldid=89656">https://halldweb.jlab.org/wiki/index.php?title=GlueX_Offline_Software_Meeting,_October_16,_2018&oldid=89656</a>"</div>
          </div>
        </div>
      </div>
      <div id="footer" role="contentinfo">
        <ul id="f-list">
          <li id="lastmod"> This page was last modified on 17 October
            2018, at 19:15.</li>
        </ul>
      </div>
    </div>
  </body>
</html>