<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    2.3 GeV data recon speed is x2 slower than the 1.05 data recon
    speed.<br>
    This is concerning, but maybe there is a good reason for that, i.e.
    might be hit occupancies are high etc,<br>
    <br>
    but the other thing, that seems puzzling, is recon speed for 2.3 GeV
    data is now more than x2 slower<br>
    using the same hps-java version, same detector, same java version,
    as we did for pass0.<br>
    <br>
    Rafo<br>
    <br>
    <br>
    <div class="moz-cite-prefix">On 05/22/2017 03:10 PM, Graham, Mathew
      Thomas wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:4AB85148-FF7F-436B-9F1D-ABC163ACD234@slac.stanford.edu">
      <div class=""><br class="">
      </div>
      <div class="">Sorry to ask a dumb question…but I will anyway.  </div>
      <div class=""><br class="">
      </div>
      <div class="">This is a “2.3 GeV vs 1.05 GeV” issue?  Or an
        hps-java version issue? </div>
      <br class="">
      <div>
        <blockquote type="cite" class="">
          <div class="">On May 22, 2017, at 12:05 PM, Omar Moreno <<a
              href="mailto:email@omarmoreno.net" class=""
              moz-do-not-send="true">email@omarmoreno.net</a>> wrote:</div>
          <br class="Apple-interchange-newline">
          <div class="">
            <div dir="auto" class="">Jeremy has also profiled the recon
              before so if Maurik can't do it, I'm sure he can. </div>
            <div class="gmail_extra"><br class="">
              <div class="gmail_quote">On May 22, 2017 11:46 AM,
                "Rafayel Paremuzyan" <<a
                  href="mailto:rafopar@jlab.org" class=""
                  moz-do-not-send="true">rafopar@jlab.org</a>> wrote:<br
                  type="attribution" class="">
                <blockquote class="quote" style="margin:0 0 0
                  .8ex;border-left:1px #ccc solid;padding-left:1ex">
                  Hi Alessandra and all,<br class="">
                  <br class="">
                  Yes clearly I see that pattern as well,<br class="">
                  going to the faster machine (pumpkin1), as Maurik
                  suggested,<br class="">
                  recon speed increased, but 2016 recon was about x2
                  slower than the 2015 recon speed.<br class="">
                  I also tried to with -Xms=4000m -Xmx=4000m options,
                  and without that options,<br class="">
                  but I didn't notice noticeable speed improvement.<br
                    class="">
                  <br class="">
                  Another thing that is *concering* is I run<br class="">
                  same jar on batch farms, centos6. and centos7, and<br
                    class="">
                  can't get same speed as we got during the pass0.<br
                    class="">
                  Note, jar is the same,  detector is the same, run
                  number and file number is the same.<br class="">
                  Difference is the time when it started to tun (Oct 19
                  2016 vs May 21 2017)<br class="">
                  <br class="">
                  If someone is interested to look at job log files<br
                    class="">
                  The log file that I run yesterday:
                  /lustre/expphy/work/hallb/hps/<wbr class="">data/physrun2016/tpass1/logs/h<wbr
                    class="">ps_008054.27_pass0XML.err<br class="">
                  The log file that run on Oct 19 2016 for pass0
                  /lustre/expphy/work/hallb/hps/<wbr class="">data/physrun2016/pass0/logs/hp<wbr
                    class="">s_008054.27_R3.9.err<br class="">
                  <br class="">
                  I have never done profiling with java.<br class="">
                  I know Maurik is in workshop, he might not have time
                  to do this,<br class="">
                  if someone else is set up doing this, it will probably
                  be useful.<br class="">
                  <br class="">
                  Also, I noticed /group/hps is full<br class="">
                  <br class="">
                  Has anyone recently put some data there?<br class="">
                  <br class="">
                  Rafo
                  <div class="elided-text"><br class="">
                    <br class="">
                    <br class="">
                    On 05/22/2017 04:04 AM, Alessandra Filippi wrote:<br
                      class="">
                  </div>
                  <blockquote class="gmail_quote" style="margin:0 0 0
                    .8ex;border-left:1px #ccc solid;padding-left:1ex">
                    <div class="elided-text">Hi Maurik, Rafo and all,<br
                        class="">
                      I think that the different Java VM could impact on
                      the absolute speed comparing different machines
                      and jar versions compiled in different times...
                      but I am running the same jar (hps-java 3.10,
                      compiled afresh at slac on rhel6-64, jvm 1.7.0) on
                      2015 and 2016 data, and the factor ~2 of speed
                      decrease for newer data is striking (whichever
                      geometry version).<br class="">
                      About garbage collection, I don't use any flag in
                      compilation, so it acts in default mode in both
                      cases.<br class="">
                      <br class="">
                      Are 2016 data affected by more noisy hits that
                      could extend the reconstruction time over testing
                      all different strategies, to your knowledge?<br
                        class="">
                      cheers<br class="">
                          Alessandra<br class="">
                      <br class="">
                      <br class="">
                      <br class="">
                      <br class="">
                      <br class="">
                      <br class="">
                      On Sun, 21 May 2017, Maurik Holtrop wrote:<br
                        class="">
                      <br class="">
                    </div>
                    <blockquote class="gmail_quote" style="margin:0 0 0
                      .8ex;border-left:1px #ccc solid;padding-left:1ex">
                      <div class="elided-text">Hello Rafo,<br class="">
                        One thing that probably is different between the
                        last time we ran with the 3.8 jar and now<br
                          class="">
                        is a different version of the Java VM. It could
                        well be that the newer version of Java is<br
                          class="">
                        not faster. Also, it is tricky to compare
                        Endeavour with the Jlab farm computers. They are<br
                          class="">
                        probably not equivalent in speed. At UNH,
                        Pumpkin has the more modern processors, whereas<br
                          class="">
                        Endeavour is now ~5 years old.<br class="">
                        <br class="">
                        Best,<br class="">
                        Maurik<br class="">
                        <br class="">
                              On May 21, 2017, at 6:54 AM, Rafayel
                        Paremuzyan <<a href="mailto:rafopar@jlab.org"
                          target="_blank" class=""
                          moz-do-not-send="true">rafopar@jlab.org</a>>
                        wrote:<br class="">
                        <br class="">
                        Hi Alessandra, Norman, all<br class="">
                        <br class="">
                        thank you for replay and your tests.<br class="">
                        <br class="">
                        I tested both 2015 and 2016 data using v4-4
                        detector on UNH computers.<br class="">
                        I have use 3.8 JAR (the jar for 2015 pass6). 3.9
                        JAR (the jar for 2016 pass0 recon),<br class="">
                        and the NEW jar v051717 (the newest jar tag is
                        v051717)<br class="">
                        <br class="">
                        Ok, I also noticed that recon of 2015 data is
                        faster that 2016 data.<br class="">
                        Also seems the new jar is 20% slower than the
                        3.9 jar for 2016 data, and about 60%<br class="">
                        slower for 2015 data.<br class="">
                        now recon speed is about 2.55 EV/S for 2015
                        data, This is too slow<br class="">
                        it cause more than 40h for a single file.<br
                          class="">
                        <br class="">
                        Ths is summary of code speeds with different jar
                        files<br class="">
                        V4-4 Detector, UNH (endeavour), 5K events are
                        reconstructed<br class="">
                        <br class="">
                        Events per second<br class="">
                        Events per second<br class="">
                        Events per second<br class="">
                        <br class="">
                        3.8JAR (2015 recon jar)<br class="">
                        3.9JAR, 2016 pass0 recon jar<br class="">
                        v051717JAR, jar vor tpass1<br class="">
                        2015 Data 5772, file 20<br class="">
                        5.07<br class="">
                        5.19<br class="">
                        3.157<br class="">
                        2016 Data file 25<br class="">
                        <br class="">
                        3.11<br class="">
                        2.53<br class="">
                        <br class="">
                        *However* I looked into job Wall times for pass0
                        recon.<br class="">
                        The recon speed is more than 7.4 Events/sec,
                        which is about x3 faster than with the<br
                          class="">
                        new JAR.<br class="">
                        <br class="">
                        I again checked *same 3.9 jar*, which is slower
                        again.<br class="">
                        I don't know why now the code speed is so low!<br
                          class="">
                        <br class="">
                        <br class="">
                        Norman, I have tried the
                        "-DdisableSvtAlignmentConstant<wbr class="">s",
                        but it didn't work<br class="">
                        <br class="">
                        =================The command===============<br
                          class="">
                        java -XX:+UseSerialGC -cp
                        hps-distribution-3.11-v051717-<wbr class="">bin.jar<br
                          class="">
                        org.hps.evio.EvioToLcio -x
                        /org/hps/steering/recon/Physic<wbr class="">sRun2016FullRecon.lcsim
                        -r<br class="">
                        -d HPS-PhysicsRun2016-v5-3-fieldm<wbr class="">ap_globalAlign
                        -R 7796 -DoutputFile=out_7796_0<br class="">
                        -DdisableSvtAlignmentConstants
                        hps_007796.evio.25 -n 10000<br class="">
                        <br class="">
                        ============The error backtrache============<br
                          class="">
                        2017-05-21 00:45:39 [CONFIG]
                        org.hps.evio.EvioToLcio parse :: using steering<br
                          class="">
                        resource /org/hps/steering/recon/Physic<wbr
                          class="">sRun2016FullRecon.lcsim<br class="">
                        2017-05-21 00:45:39 [CONFIG]
                        org.hps.evio.EvioToLcio parse :: set max events
                        to<br class="">
                        10000<br class="">
                        2017-05-21 00:45:48 [INFO]
                        org.hps.rundb.RunManager <init> ::
                        ConnectionParameters {<br class="">
                        database: hps_run_db_v2, hostname: <a
                          href="http://hpsdb.jlab.org/" rel="noreferrer"
                          target="_blank" class=""
                          moz-do-not-send="true">
                          hpsdb.jlab.org</a>, password: darkphoton,
                        port: 3306,<br class="">
                        user: hpsuser }<br class="">
                        2017-05-21 00:45:48 [CONFIG]
                        org.lcsim.job.JobControlManage<wbr class="">r
                        addVariableDefinition<br class="">
                        :: outputFile = out_7796_0<br class="">
                        2017-05-21 00:45:48 [CONFIG]
                        org.hps.evio.EvioToLcio parse :: set steering
                        variable:<br class="">
                        outputFile=out_7796_0<br class="">
                        2017-05-21 00:45:48 [SEVERE]
                        org.hps.evio.EvioToLcio parse :: bad variable
                        format:<br class="">
                        disableSvtAlignmentConstants<br class="">
                        java.lang.IllegalArgumentExcep<wbr class="">tion:
                        Bad variable format:<br class="">
                        disableSvtAlignmentConstants<br class="">
                                at org.hps.evio.EvioToLcio.parse(<wbr
                          class="">EvioToLcio.java:393)<br class="">
                                at org.hps.evio.EvioToLcio.main(E<wbr
                          class="">vioToLcio.java:97)<br class="">
                        <br class="">
                        Exception in thread "main"
                        java.lang.IllegalArgumentExcep<wbr class="">tion:
                        Bad variable format:<br class="">
                        disableSvtAlignmentConstants<br class="">
                                at org.hps.evio.EvioToLcio.parse(<wbr
                          class="">EvioToLcio.java:393)<br class="">
                                at org.hps.evio.EvioToLcio.main(E<wbr
                          class="">vioToLcio.java:97)<br class="">
                        <br class="">
                        Rafo<br class="">
                        <br class="">
                        <br class="">
                        <br class="">
                        <br class="">
                        <br class="">
                        <br class="">
                        On 05/20/2017 06:17 AM, Alessandra Filippi
                        wrote:<br class="">
                              Hi Rafo, all,<br class="">
                              I also noticed that the reconstruction for
                        2016 data is about twice as<br class="">
                              slow as compared to 2015 (whichever
                        geometry and reconstruction<br class="">
                              version).<br class="">
                              This happens when I run the aligned
                        geometry as well as the "current"<br class="">
                              one (v5.0), and the geometry taken from
                        the db as well (the result is<br class="">
                              the same as v5.0). I did not make any test
                        with v4.4, though - actually<br class="">
                              as far as svt alignment is concerned it
                        should be the same as v5.0. Can<br class="">
                              you please try and make the same short
                        test with the newest jar with<br class="">
                              v4.4?<br class="">
                              This happens to me both with hps-java 5.10
                        and 5.11 (not the most<br class="">
                              updated one).<br class="">
                        <br class="">
                              I would be surprised if it could be
                        something connected to the<br class="">
                              alignment, unless for some reason new
                        positions and harder tracks<br class="">
                              trigger some long loops in the
                        reconstruction. But this happens (to me)<br
                          class="">
                              also with the<br class="">
                              standard geometry, so a check with the one
                        used with pass0 (that should<br class="">
                              however be equivalent to v5.0) could at
                        least help to rule out, or blame<br class="">
                              on, the alignment step.<br class="">
                              Thanks, cheers<br class="">
                                  Alessandra<br class="">
                        <br class="">
                        <br class="">
                              ps. make also sure that the correct
                        fieldmap is called in all the<br class="">
                              compact files - you never know!<br
                          class="">
                        <br class="">
                        <br class="">
                        <br class="">
                              On Fri, 19 May 2017, Rafayel Paremuzyan
                        wrote:<br class="">
                        <br class="">
                                    Hi All,<br class="">
                        <br class="">
                                    During the testing the recon for
                        test pass1,<br class="">
                                    I noticed the recon time is more
                        than x2 longer wrt pass0<br class="">
                                    recon time.<br class="">
                        <br class="">
                                    To demonstrate it<br class="">
                                    I submit 3 simple jobs with 10K
                        events to reconstruct, with<br class="">
                                    new pass1 xml<br class="">
                                    file (this has the new jar v051717,
                        and the new detector<br class="">
                                    HPS-PhysicsRun2016-v5-3-fieldm<wbr
                          class="">ap_globalAlign),<br class="">
                                    and the old pass0 xml file (pass0
                        jar release 3.9, and the<br class="">
                                    detector<br class="">
                                    HPS-PhysicsRun2016-Nominal-v4-<wbr
                          class="">4-fieldmap)<br class="">
                        <br class="">
                                    Below is a printout from the jobs
                        with a new JAR, v051717,<br class="">
                                    the average time<br class="">
                                    for 1000 events is more than 7
                        minutes<br class="">
                                    ===================== LOG from the
                        v051717 JAR<br class="">
                                    ==============================<br
                          class="">
                                    2017-05-19 09:36:51 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10614074 with sequence 0<br class="">
                                    2017-05-19 09:43:13 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10615074 with sequence 1000<br
                          class="">
                                    2017-05-19 09:49:18 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10616074 with sequence 2000<br
                          class="">
                                    2017-05-19 09:55:54 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10617074 with sequence 3000<br
                          class="">
                                    2017-05-19 10:02:55 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10618074 with sequence 4000<br
                          class="">
                                    2017-05-19 10:09:57 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10619074 with sequence 5000<br
                          class="">
                                    2017-05-19 10:16:13 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10620074 with sequence 6000<br
                          class="">
                                    2017-05-19 10:25:20 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10621074 with sequence 7000<br
                          class="">
                                    2017-05-19 10:32:56 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10622074 with sequence 8000<br
                          class="">
                                    2017-05-19 10:36:19 [WARNING]<br
                          class="">
                                    org.hps.recon.tracking.Tracker<wbr
                          class="">ReconDriver<br class="">
                                    process :: Discarding track with bad
                        HelicalTrackHit<br class="">
                                    (correction distance<br class="">
                                    0.000000, chisq penalty 0.000000)<br
                          class="">
                                    2017-05-19 10:42:03 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10623074 with sequence 9000<br
                          class="">
                                    2017-05-19 10:47:44 [INFO]
                        org.hps.evio.EvioToLcio run ::<br class="">
                                    maxEvents 10000<br class="">
                                    was reached<br class="">
                                    2017-05-19 10:47:44 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    endOfData ::<br class="">
                                    10000 events processed in job.<br
                          class="">
                                    2017-05-19 10:47:44 [INFO]
                        org.hps.evio.EvioToLcio run ::<br class="">
                                    Job finished<br class="">
                                    successfully!<br class="">
                        <br class="">
                        <br class="">
                                    And below is the Job log info from
                        the pass0 jar. The<br class="">
                                    average time for 1000<br class="">
                                    events is less than 3 minutes<br
                          class="">
                                    ===================== LOG from the
                        3.9 release JAR<br class="">
                                    ==============================<br
                          class="">
                                    2017-05-19 13:19:46 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10614074 with sequence 0<br class="">
                                    2017-05-19 13:23:36 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10615074 with sequence 1000<br
                          class="">
                                    2017-05-19 13:27:03 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10616074 with sequence 2000<br
                          class="">
                                    2017-05-19 13:30:40 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10617074 with sequence 3000<br
                          class="">
                                    2017-05-19 13:34:20 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10618074 with sequence 4000<br
                          class="">
                                    2017-05-19 13:38:11 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10619074 with sequence 5000<br
                          class="">
                                    2017-05-19 13:41:43 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10620074 with sequence 6000<br
                          class="">
                                    2017-05-19 13:45:54 [WARNING]<br
                          class="">
                                    org.hps.recon.tracking.Tracker<wbr
                          class="">ReconDriver<br class="">
                                    process :: Discarding track with bad
                        HelicalTrackHit<br class="">
                                    (correction distance<br class="">
                                    0.000000, chisq penalty 0.000000)<br
                          class="">
                                    2017-05-19 13:46:05 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10621074 with sequence 7000<br
                          class="">
                                    2017-05-19 13:50:08 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10622074 with sequence 8000<br
                          class="">
                                    2017-05-19 13:55:03 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    process :: Event<br class="">
                                    10623074 with sequence 9000<br
                          class="">
                                    2017-05-19 13:58:27 [INFO]
                        org.hps.evio.EvioToLcio run ::<br class="">
                                    maxEvents 10000<br class="">
                                    was reached<br class="">
                                    2017-05-19 13:58:27 [INFO]
                        org.lcsim.job.EventMarkerDrive<wbr class="">r<br
                          class="">
                                    endOfData ::<br class="">
                                    10000 events processed in job.<br
                          class="">
                                    2017-05-19 13:58:27 [INFO]
                        org.hps.evio.EvioToLcio run ::<br class="">
                                    Job finished<br class="">
                                    successfully!<br class="">
                        <br class="">
                                    I also tried to do reconstruction by
                        myself interactively,<br class="">
                                    but I am getting<br class="">
                                    error below,<br class="">
                        <br class="">
                                    The command<br class="">
                                    /apps/scicomp/java/jdk1.7/bin/<wbr
                          class="">java -XX:+UseSerialGC -cp<br class="">
                                    hps-distribution-3.9-bin.jar
                        org.hps.evio.EvioToLcio -x<br class="">
                        /org/hps/steering/recon/Physic<wbr class="">sRun2016FullRecon.lcsim
                        -r -d<br class="">
                                    HPS-PhysicsRun2016-v5-3-fieldm<wbr
                          class="">ap_globalAlign -R 7796<br class="">
                                    -DoutputFile=out_7796_0<br class="">
                                    hps_007796.evio.0 -n 10000<br
                          class="">
                        <br class="">
                                    The Error traceback<br class="">
                                    017-05-19 14:58:44 [CONFIG]
                        org.hps.evio.EvioToLcio parse ::<br class="">
                                    using steering<br class="">
                                    resource<br class="">
                        /org/hps/steering/recon/Physic<wbr class="">sRun2016FullRecon.lcsim<br
                          class="">
                                    2017-05-19 14:58:44 [CONFIG]
                        org.hps.evio.EvioToLcio parse<br class="">
                                    :: set max events<br class="">
                                    to 10000<br class="">
                                    2017-05-19 14:58:45 [CONFIG]<br
                          class="">
                                    org.lcsim.job.JobControlManage<wbr
                          class="">r<br class="">
                                    addVariableDefinition :: outputFile
                        = out_7796_0<br class="">
                                    2017-05-19 14:58:45 [CONFIG]
                        org.hps.evio.EvioToLcio parse<br class="">
                                    :: set steering<br class="">
                                    variable: outputFile=out_7796_0<br
                          class="">
                                    2017-05-19 14:58:45 [CONFIG]
                        org.lcsim.job.JobControlManage<wbr class="">r<br
                          class="">
                                    initializeLoop<br class="">
                                    :: initializing LCSim loop<br
                          class="">
                                    2017-05-19 14:58:45 [CONFIG]
                        org.lcsim.job.JobControlManage<wbr class="">r<br
                          class="">
                                    initializeLoop<br class="">
                                    :: Event marker printing disabled.<br
                          class="">
                                    2017-05-19 14:58:45 [INFO]<br
                          class="">
                        org.hps.conditions.database.Da<wbr class="">tabaseConditionsManager<br
                          class="">
                                    resetInstance ::<br class="">
                                    DatabaseConditionsManager instance
                        is reset<br class="">
                                    Exception in thread "main"
                        java.lang.UnsatisfiedLinkError<wbr class="">:<br
                          class="">
                        /u/apps/scicomp/java/jdk1.7.0_<wbr class="">75/jre/lib/i386/xawt/libmawt.s<wbr
                          class="">o:<br class="">
                                    libXext.so.6:<br class="">
                                    cannot open shared object file: No
                        such file or directory<br class="">
                                            at
                        java.lang.ClassLoader$NativeLi<wbr class="">brary.load(Native<br
                          class="">
                                    Method)<br class="">
                                            at<br class="">
                        java.lang.ClassLoader.loadLibr<wbr class="">ary1(ClassLoader.java:1965)<br
                          class="">
                                            at<br class="">
                        java.lang.ClassLoader.loadLibr<wbr class="">ary0(ClassLoader.java:1890)<br
                          class="">
                                            at<br class="">
                        java.lang.ClassLoader.loadLibr<wbr class="">ary(ClassLoader.java:1851)<br
                          class="">
                                            at
                        java.lang.Runtime.load0(Runtim<wbr class="">e.java:795)<br
                          class="">
                                            at
                        java.lang.System.load(System.j<wbr class="">ava:1062)<br
                          class="">
                                            at
                        java.lang.ClassLoader$NativeLi<wbr class="">brary.load(Native<br
                          class="">
                                    Method)<br class="">
                                            at<br class="">
                        java.lang.ClassLoader.loadLibr<wbr class="">ary1(ClassLoader.java:1965)<br
                          class="">
                                            at<br class="">
                        java.lang.ClassLoader.loadLibr<wbr class="">ary0(ClassLoader.java:1890)<br
                          class="">
                                            at<br class="">
                        java.lang.ClassLoader.loadLibr<wbr class="">ary(ClassLoader.java:1872)<br
                          class="">
                                            at
                        java.lang.Runtime.loadLibrary0<wbr class="">(Runtime.java:849)<br
                          class="">
                                            at
                        java.lang.System.loadLibrary(S<wbr class="">ystem.java:1088)<br
                          class="">
                                            at<br class="">
                        sun.security.action.LoadLibrar<wbr class="">yAction.run(LoadLibraryAction.<wbr
                          class="">java:67)<br class="">
                                            at<br class="">
                        sun.security.action.LoadLibrar<wbr class="">yAction.run(LoadLibraryAction.<wbr
                          class="">java:47)<br class="">
                                            at<br class="">
                                    java.security.AccessController<wbr
                          class="">.doPrivileged(Native Method)<br
                          class="">
                                            at<br class="">
                                    java.awt.Toolkit.loadLibraries<wbr
                          class="">(Toolkit.java:1653)<br class="">
                                            at
                        java.awt.Toolkit.<clinit>(Tool<wbr
                          class="">kit.java:1682)<br class="">
                                            at
                        java.awt.Component.<clinit>(Co<wbr
                          class="">mponent.java:595)<br class="">
                                            at
                        org.lcsim.util.aida.AIDA.<init<wbr class="">>(AIDA.java:68)<br
                          class="">
                                            at<br class="">
                        org.lcsim.util.aida.AIDA.defau<wbr class="">ltInstance(AIDA.java:53)<br
                          class="">
                                            at<br class="">
                        org.hps.evio.RfFitterDriver.<i<wbr class="">nit>(RfFitterDriver.java:31)<br
                          class="">
                                            at<br class="">
                        sun.reflect.NativeConstructorA<wbr class="">ccessorImpl.newInstance0(<wbr
                          class="">Native<br class="">
                                    Method)<br class="">
                                atsun.reflect.NativeConstructo<wbr
                          class="">rAccessorImpl.newInstance(Nati<wbr
                          class="">veConstructorAcce<br class="">
                                    ssorImpl.java:57)<br class="">
                                atsun.reflect.DelegatingConstr<wbr
                          class="">uctorAccessorImpl.newInstance(<wbr
                          class="">DelegatingConstru<br class="">
                                    ctorAccessorImpl.java:45)<br
                          class="">
                                            at<br class="">
                        java.lang.reflect.Constructor.<wbr class="">newInstance(Constructor.java:5<wbr
                          class="">26)<br class="">
                                            at
                        java.lang.Class.newInstance(Cl<wbr class="">ass.java:379)<br
                          class="">
                                            at<br class="">
                        org.lcsim.job.JobControlManage<wbr class="">r.setupDrivers(JobControlManag<wbr
                          class="">er.java:1199)
                        <br class="">
                                            at<br class="">
                        org.hps.job.JobManager.setupDr<wbr class="">ivers(JobManager.java:82)<br
                          class="">
                                            at<br class="">
                        org.lcsim.job.JobControlManage<wbr class="">r.setup(JobControlManager.<wbr
                          class="">java:1052)<br class="">
                                            at<br class="">
                        org.lcsim.job.JobControlManage<wbr class="">r.setup(JobControlManager.<wbr
                          class="">java:1110)<br class="">
                                            at<br class="">
                                    org.hps.evio.EvioToLcio.parse(<wbr
                          class="">EvioToLcio.java:407)<br class="">
                                            at
                        org.hps.evio.EvioToLcio.main(E<wbr class="">vioToLcio.java:97)<br
                          class="">
                        <br class="">
                        <br class="">
                        <br class="">
                                    I see this library libXext.so.6: in
                        /usr/lib64, but not in<br class="">
                                    /usr/lib,<br class="">
                                    when I put /usr/lib64 in my
                        LD_LIBRARY_PATH, then it<br class="">
                                    complaines again (see<br class="">
                                    below)<br class="">
                        <br class="">
                                    Exception in thread "main"
                        java.lang.UnsatisfiedLinkError<wbr class="">:<br
                          class="">
                        /u/apps/scicomp/java/jdk1.7.0_<wbr class="">75/jre/lib/i386/xawt/libmawt.s<wbr
                          class="">o:<br class="">
                                    libXext.so.6:<br class="">
                                    wrong ELF class: ELFCLASS64<br
                          class="">
                        <br class="">
                                    I would appreciate, if I get some
                        help on running the<br class="">
                                    reconstruction<br class="">
                                    interactively, then I could look
                        more closely into logs<br class="">
                                    of the old, and new JAR files.<br
                          class="">
                        <br class="">
                                    Rafo<br class="">
                        <br class="">
                        <br class="">
                        ______________________________<wbr class="">______________________________<wbr
                          class="">________________
                        <br class="">
                        <br class="">
                                    Use REPLY-ALL to reply to list<br
                          class="">
                        <br class="">
                                    To unsubscribe from the HPS-SOFTWARE
                        list, click the<br class="">
                                    following link:<br class="">
                        <a
href="https://urldefense.proofpoint.com/v2/url?u=https-3A__listserv.slac.stanford.edu_cgi-2Dbin"
                          rel="noreferrer" target="_blank" class=""
                          moz-do-not-send="true">https://urldefense.proofpoint.<wbr
                            class="">com/v2/url?u=https-3A__listser<wbr
                            class="">v.slac.stanford.edu_cgi-2Dbin</a>
                        <br class="">
                        _wa-3FSUBED1-3DHPS-2DSOFTWARE-<wbr class="">26A-3D1&d=DwIDaQ&c=lz9TcOasaIN<wbr
                          class="">aaC3U7FbMev2lsutwpI4--09aP8Lu
                        <br class="">
                        18s&r=0HDJrGO9TZQTE97J9Abt2A&m<wbr
                          class="">=xnbGP5VHYWRAQRWRksVgMnYvBkXWI<wbr
                          class="">4roLxztdJ0Tp9I&s=ppNYedSrn5DP
                        <br class="">
                                    aIZZJgRZu8tBDeSjroqbj_PoevFoFp<wbr
                          class="">I&e=<br class="">
                        <br class="">
                        <br class="">
                        <br class="">
                        ##############################<wbr class="">##############################<wbr
                          class="">############<br class="">
                              Use REPLY-ALL to reply to list<br class="">
                        <br class="">
                              To unsubscribe from the HPS-SOFTWARE list,
                        click the following link:<br class="">
                        <a
href="https://urldefense.proofpoint.com/v2/url?u=https-3A__listserv.slac.stanford.edu_cgi-2Dbin"
                          rel="noreferrer" target="_blank" class=""
                          moz-do-not-send="true">https://urldefense.proofpoint.<wbr
                            class="">com/v2/url?u=https-3A__listser<wbr
                            class="">v.slac.stanford.edu_cgi-2Dbin</a>
                        <br class="">
                        _wa-3FSUBED1-3DHPS-2DSOFTWARE-<wbr class="">26A-3D1&d=DwIDaQ&c=lz9TcOasaIN<wbr
                          class="">aaC3U7FbMev2lsutwpI4--09aP8Lu
                        <br class="">
                        18s&r=0HDJrGO9TZQTE97J9Abt2A&m<wbr
                          class="">=xnbGP5VHYWRAQRWRksVgMnYvBkXWI<wbr
                          class="">4roLxztdJ0Tp9I&s=ppNYedSrn5DP
                        <br class="">
                              aIZZJgRZu8tBDeSjroqbj_PoevFoFp<wbr
                          class="">I&e=<br class="">
                        <br class="">
                        <br class="">
                        <br class="">
                        ______________________________<wbr class="">______________________________<wbr
                          class="">______________________________
                        <br class="">
                        <br class="">
                        Use REPLY-ALL to reply to list<br class="">
                        <br class="">
                        To unsubscribe from the HPS-SOFTWARE list, click
                        the following link:<br class="">
                      </div>
                      <a
href="https://urldefense.proofpoint.com/v2/url?u=https-3A__listserv.slac.stanford.edu_cgi-2Dbin_wa-3FSUBED1-3DHPS-2DSOFTWARE-26A-3D1&d=DwIDbA&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=0HDJrGO9TZQTE97J9Abt2A&m=nIEvNWJSNxvmECWKfMrVKMTZKNdym3zsAls7WG1fdfU&s=cI82nYRtYBzgLs1l5g3anHrSt0kPk4B98oUzllu5KfQ&e="
                        rel="noreferrer" target="_blank" class=""
                        moz-do-not-send="true">https://urldefense.proofpoint.<wbr
                          class="">com/v2/url?u=https-3A__listser<wbr
                          class="">v.slac.stanford.edu_cgi-2Dbin_<wbr
                          class="">wa-3FSUBED1-3DHPS-2DSOFTWARE-<wbr
                          class="">26A-3D1&d=DwIDbA&c=lz9TcOasaIN<wbr
                          class="">aaC3U7FbMev2lsutwpI4--<wbr class="">09aP8Lu18s&r=0HDJrGO9TZQTE97J9<wbr
                          class="">Abt2A&m=nIEvNWJSNxvmECWKfMrVKM<wbr
                          class="">TZKNdym3zsAls7WG1fdfU&s=cI82nY<wbr
                          class="">RtYBzgLs1l5g3anHrSt0kPk4B98oUz<wbr
                          class="">llu5KfQ&e=</a>
                      <br class="">
                      <br class="">
                      <br class="">
                      <br class="">
                    </blockquote>
                    <div class="quoted-text"><br class="">
                      ##############################<wbr class="">##############################<wbr
                        class="">############<br class="">
                      Use REPLY-ALL to reply to list<br class="">
                      <br class="">
                      To unsubscribe from the HPS-SOFTWARE list, click
                      the following link:<br class="">
                    </div>
                    <a
href="https://urldefense.proofpoint.com/v2/url?u=https-3A__listserv.slac.stanford.edu_cgi-2Dbin_wa-3FSUBED1-3DHPS-2DSOFTWARE-26A-3D1&d=DwIDbA&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=0HDJrGO9TZQTE97J9Abt2A&m=nIEvNWJSNxvmECWKfMrVKMTZKNdym3zsAls7WG1fdfU&s=cI82nYRtYBzgLs1l5g3anHrSt0kPk4B98oUzllu5KfQ&e="
                      rel="noreferrer" target="_blank" class=""
                      moz-do-not-send="true">https://urldefense.proofpoint.<wbr
                        class="">com/v2/url?u=https-3A__listser<wbr
                        class="">v.slac.stanford.edu_cgi-2Dbin_<wbr
                        class="">wa-3FSUBED1-3DHPS-2DSOFTWARE-<wbr
                        class="">26A-3D1&d=DwIDbA&c=lz9TcOasaIN<wbr
                        class="">aaC3U7FbMev2lsutwpI4--<wbr class="">09aP8Lu18s&r=0HDJrGO9TZQTE97J9<wbr
                        class="">Abt2A&m=nIEvNWJSNxvmECWKfMrVKM<wbr
                        class="">TZKNdym3zsAls7WG1fdfU&s=cI82nY<wbr
                        class="">RtYBzgLs1l5g3anHrSt0kPk4B98oUz<wbr
                        class="">llu5KfQ&e=</a>
                    <br class="">
                  </blockquote>
                  <br class="">
                  <br class="">
                  ______________________________<wbr class="">_________________<br
                    class="">
                  Hps-analysis mailing list<br class="">
                  <a href="mailto:Hps-analysis@jlab.org" target="_blank"
                    class="" moz-do-not-send="true">Hps-analysis@jlab.org</a><br
                    class="">
                  <a
                    href="https://mailman.jlab.org/mailman/listinfo/hps-analysis"
                    rel="noreferrer" target="_blank" class=""
                    moz-do-not-send="true">https://mailman.jlab.org/mailm<wbr
                      class="">an/listinfo/hps-analysis</a><br class="">
                </blockquote>
              </div>
              <br class="">
            </div>
            _______________________________________________<br class="">
            Hps-analysis mailing list<br class="">
            <a href="mailto:Hps-analysis@jlab.org" class=""
              moz-do-not-send="true">Hps-analysis@jlab.org</a><br
              class="">
            <a class="moz-txt-link-freetext" href="https://mailman.jlab.org/mailman/listinfo/hps-analysis">https://mailman.jlab.org/mailman/listinfo/hps-analysis</a><br
              class="">
          </div>
        </blockquote>
      </div>
      <br class="">
      <br>
      <hr>
      <p align="left">
        Use REPLY-ALL to reply to list</p>
      <p align="center">To unsubscribe from the HPS-SOFTWARE list, click
        the following link:<br>
        <a
href="https://urldefense.proofpoint.com/v2/url?u=https-3A__listserv.slac.stanford.edu_cgi-2Dbin_wa-3FSUBED1-3DHPS-2DSOFTWARE-26A-3D1&d=DwMGaQ&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=0HDJrGO9TZQTE97J9Abt2A&m=yktqxWZIxrOJDj0GcxtXMdIm5S6CzoD5o_bfENc8yOA&s=9UGWhMTkreQfG7iEUol-dsVGExrJm-lRdRBfTmh68jU&e="
          target="_blank" moz-do-not-send="true">https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1</a>
      </p>
    </blockquote>
    <br>
  </body>
</html>