Difference between revisions of "GlueX Data Challenge Meeting, October 8, 2012"

From GlueXWiki
Jump to: navigation, search
(Agenda)
 
(6 intermediate revisions by 2 users not shown)
Line 28: Line 28:
 
#** Instructions for the [http://evo.caltech.edu/evoGate/Documentation/extclient/telephone/telephone.html Phone Bridge] to EVO.
 
#** Instructions for the [http://evo.caltech.edu/evoGate/Documentation/extclient/telephone/telephone.html Phone Bridge] to EVO.
 
#* [http://evo.caltech.edu/evoGate/Documentation/extclient/skype/skype.html Skype Bridge] to EVO
 
#* [http://evo.caltech.edu/evoGate/Documentation/extclient/skype/skype.html Skype Bridge] to EVO
 +
 +
=Minutes=
 +
 +
Present:
 +
* '''CMU''': Paul Mattione, Curtis Meyer
 +
* '''JLab''': Mark Ito (chair), David Lawrence, Simon Taylor
 +
 +
==JLab Farm Priority==
 +
 +
Paul asked if there was a special account. one that gets increased priority, that we could use, like the production accounts in CLAS (e. g., clasg13). Mark noted that those accounts are enhanced only relative to other CLAS users. There is still an overall allocation per Hall. Hall D's allocation is currently 4% no matter how we split it up internally.
 +
 +
More broadly, Mark promised to research how the current scheme is executed in detail. His memory of these details are from the old LSF days.
 +
 +
==JLab Farm Mini-Challenge Status==
 +
 +
Mark showed the [[Mini Data Challenge Status, October 8, 2012|summary statistics]] from the now-completed mini-challenge which was producing REST formatted data. There is still a poor yield (79%) of output files produced by a job, and that in fact not all of those were complete. The outstanding memory issue has not been addressed, however.
 +
 +
==Large, Usable Pythia Sample==
 +
 +
We discussed the idea, endorsed at the collaboration meeting, of making an effort to generate 2 billion Pythia events on some combination of our current resources. Mark [[Large, usable Pythia sample|outlined some of the issues]]. We also discussed [[Software improvements from Richard|Richard's recent email]] describing software changes he advocates before we get started. Some points from the discussion:
 +
 +
* We need a standard software suite before starting. Practically that means putting in things that are not in the repository yet, and eliminating some things that are. We decided:
 +
** to put in the changes to REST format that Richard mentioned
 +
** to not wait for the changes to the HDDM format that he describes since the HDDM_s data will not be kept anyway (for the most part). We will have to run this argument by Richard.
 +
** to not include the new BCAL digitization scheme from David
 +
* We need to standardize the configuration files, command line arguments, and random number seed schemes across all production sites.
 +
* We will start at all three sites and see how things go, i. e., no per-site goal at first.
 +
* We thought that 2 billion events was indeed a reasonable total to shoot for.
 +
* We set two weeks (the next meeting) as a goal for being ready to start.
 +
* We enjoined Curtis to make a request to IT Division to increase our share of the farm to 50% from 4%. (He actually volunteered.)

Latest revision as of 16:24, 9 October 2012

GlueX Data Challenge Meeting
Monday, October 8, 2012
1:30 pm, EDT
JLab: CEBAF Center, F326/327

Agenda

Meeting Connections

To connect from the outside:

  1. ESNET:
    • Call ESNET Number 8542553 (this is the preferred connection method).
  2. Phone: (should not be needed)
    • +1-866-740-1260 : US and Canada
    • +1-303-248-0285 : International
    • then use participant code: 3421244# (the # is needed when using the phone)
    • or www.readytalk.com
      • then type access code 3421244 into "join a meeting" (you need java plugin)
  3. EVO:
    • A conference has been booked under "GlueX" from 1:00pm until 3:30pm (EST).
    • Direct meeting link
    • To phone into an EVO meeting, from the U.S. call (626) 395-2112 and then enter the EVO meeting code, 13 9993
    • Skype Bridge to EVO

Minutes

Present:

  • CMU: Paul Mattione, Curtis Meyer
  • JLab: Mark Ito (chair), David Lawrence, Simon Taylor

JLab Farm Priority

Paul asked if there was a special account. one that gets increased priority, that we could use, like the production accounts in CLAS (e. g., clasg13). Mark noted that those accounts are enhanced only relative to other CLAS users. There is still an overall allocation per Hall. Hall D's allocation is currently 4% no matter how we split it up internally.

More broadly, Mark promised to research how the current scheme is executed in detail. His memory of these details are from the old LSF days.

JLab Farm Mini-Challenge Status

Mark showed the summary statistics from the now-completed mini-challenge which was producing REST formatted data. There is still a poor yield (79%) of output files produced by a job, and that in fact not all of those were complete. The outstanding memory issue has not been addressed, however.

Large, Usable Pythia Sample

We discussed the idea, endorsed at the collaboration meeting, of making an effort to generate 2 billion Pythia events on some combination of our current resources. Mark outlined some of the issues. We also discussed Richard's recent email describing software changes he advocates before we get started. Some points from the discussion:

  • We need a standard software suite before starting. Practically that means putting in things that are not in the repository yet, and eliminating some things that are. We decided:
    • to put in the changes to REST format that Richard mentioned
    • to not wait for the changes to the HDDM format that he describes since the HDDM_s data will not be kept anyway (for the most part). We will have to run this argument by Richard.
    • to not include the new BCAL digitization scheme from David
  • We need to standardize the configuration files, command line arguments, and random number seed schemes across all production sites.
  • We will start at all three sites and see how things go, i. e., no per-site goal at first.
  • We thought that 2 billion events was indeed a reasonable total to shoot for.
  • We set two weeks (the next meeting) as a goal for being ready to start.
  • We enjoined Curtis to make a request to IT Division to increase our share of the farm to 50% from 4%. (He actually volunteered.)