Difference between revisions of "GlueX Offline Meeting, August 21, 2013"

From GlueXWiki
Jump to: navigation, search
(Created page with "GlueX Offline Software Meeting<br> Wednesday, August 21, 2013<br> 1:30 pm EDT<br> JLab: CEBAF Center F326/327 =Agenda= # Announcements # Review of [[GlueX Offline Meeting, Augu...")
 
m (Text replacement - "www/halldweb1/" to "www/halldweb/")
 
(7 intermediate revisions by 2 users not shown)
Line 8: Line 8:
 
# Announcements
 
# Announcements
 
# Review of [[GlueX Offline Meeting, August 7, 2013#Minutes|minutes from the last meeting]]: all
 
# Review of [[GlueX Offline Meeting, August 7, 2013#Minutes|minutes from the last meeting]]: all
 +
# ODU Data Mining Project (Gagik)
 
# Raw Data Factories (David)
 
# Raw Data Factories (David)
 +
# [[Mattione_Update_08212013 | Spiral Track Reconstruction]] (Paul)
 +
# New HDDS version? (all)
 +
# Magnetic field switch? (all)
 +
# Report from Work Flow meeting (David, Mark)
 
# Mantis Bug Tracker Review
 
# Mantis Bug Tracker Review
 
#* [http://tinyurl.com/4yjnabe Recently Resolved]
 
#* [http://tinyurl.com/4yjnabe Recently Resolved]
Line 38: Line 43:
 
==Slides==
 
==Slides==
  
Talks can be deposited in the directory <code>/group/halld/www/halldweb1/html/talks/2013-3Q</code> on the JLab CUE. This directory is accessible from the web at https://halldweb1.jlab.org/talks/2013-3Q/ .
+
Talks can be deposited in the directory <code>/group/halld/www/halldweb/html/talks/2013-3Q</code> on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2013-3Q/ .
 +
 
 +
=Minutes=
 +
 
 +
Present:
 +
 
 +
* '''CMU''': Will Levine, Paul Mattione, Curtis Meyer
 +
* '''JLab''': Mark Dalton, David Lawrence, Mark Ito (chair), Simon Taylor, Elliott Wolin, Beni Zihlmann
 +
* '''Northwestern''': Sean Dobbs
 +
* '''UConn''': Alex Barnes
 +
 
 +
==Raw Data Factories==
 +
 
 +
David explained the new layer of classes introduced to deal with he real raw data format. Up until now the lowest level objects contained information in physical units such as GeV or nanoseconds. In preparation for the online data challenge, raw integer value data from the data acquisition is being simulated and must be converted to data in physical units. The new "digi" layer will hold the raw data and the old "hit" layer will continue to contain the post-unit-conversion data. Calibrations will be applied during this conversion.
 +
 
 +
There is in fact another layer below digi, for example the DFADC250 objects, which contain information direct from the data acquisition. The event recorder will be writing data at this lowest level to the output files. The translation tables relating geographic channel information to detector channel information are used in the conversion to the "digi" objects. In this scheme the raw data written from the experiment will not depend on the translation tables.
 +
 
 +
==Spiral Track  Reconstruction==
 +
 
 +
Paul presented preliminary results on a new effort to reduce the number of ghost track produced by tracks which spiral in the CDC. See his [[Mattione Update 08212013|wiki page]] for details.
 +
 
 +
He showed several events where the current tracking finds multiple tracks from a single generated track and where the new code finds one. He also showed average multiplicity from old and new code at the track candidate level and at the time-based track level. He also plotted the distribution for the track parameters at the track candidate level (all tracks were thrown with 150 MeV/c at 90&degrees;). The fraction of times where at least one "roughly close" track was found turns out to be similar between the two codes.
 +
 
 +
The results look very promising. Although these problem tracks are not a huge fraction of the phase space, for some reactions they may contribute significantly. There is a suspicion that they caused some of the crashes seen during the data challenge. One thing that remains to be studied is the effect of the new algorithm on multi-track events; we do not want it to be <em>too</em> good at eliminating tracks!
 +
 
 +
Paul has been consulting with Simon on this work. Simon will likely take over this work at some appropriate time in the future.
 +
 
 +
==New HDDS Version==
 +
 
 +
Simon has been working on adding the cooling manifolds on the FDC and fixing some issues with the BCAL geometry. These should not affect the nightly builds much and we are not ready to issue a new tagged release of the geometry.
 +
 
 +
==Magnetic Field Switch==
 +
 
 +
David intends to generate a new magnetic field map for 1350 A and the correct polarity. We will adopt it when it is ready.
 +
 
 +
==Report from the Work Flow Meeting==
 +
 
 +
Chip Watson has convened a committee to look at work flow tools and data provenance solutions for general use at the Lab. By this he means
 +
* '''Work Flow Tools''': standard automation of large-scale, multiple job submission and tracking
 +
* '''Data Provenance''': having information on how data was processed and archiving tools necessary to re-analyze data in the future.
 +
Scientific Computing at JLab is willing to contribute resources to design and implementation of a system. The committee consists of representatives from each of the halls and from Scientific Computing. David and Mark are representing Hall D. The first meeting is next week, on Wednesday. Progress on the effort will be reported back to the collaboration as work goes forward.

Latest revision as of 11:10, 31 March 2015

GlueX Offline Software Meeting
Wednesday, August 21, 2013
1:30 pm EDT
JLab: CEBAF Center F326/327

Agenda

  1. Announcements
  2. Review of minutes from the last meeting: all
  3. ODU Data Mining Project (Gagik)
  4. Raw Data Factories (David)
  5. Spiral Track Reconstruction (Paul)
  6. New HDDS version? (all)
  7. Magnetic field switch? (all)
  8. Report from Work Flow meeting (David, Mark)
  9. Mantis Bug Tracker Review
  10. Review of recent repository activity: all

Communication Information

Video Conferencing

Desktop Sharing

You can view the computer desktop in the meeting room at JLab via the web.

  1. Go to http://esnet.readytalk.com
  2. In the "join a meeting" box enter the Hall D code: 1833622
  3. Fill in the participant registration form.

Telephone

To connect by telephone:

  1. dial:
  2. enter access code followed by the # sign: 1833622#

Slides

Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2013-3Q on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2013-3Q/ .

Minutes

Present:

  • CMU: Will Levine, Paul Mattione, Curtis Meyer
  • JLab: Mark Dalton, David Lawrence, Mark Ito (chair), Simon Taylor, Elliott Wolin, Beni Zihlmann
  • Northwestern: Sean Dobbs
  • UConn: Alex Barnes

Raw Data Factories

David explained the new layer of classes introduced to deal with he real raw data format. Up until now the lowest level objects contained information in physical units such as GeV or nanoseconds. In preparation for the online data challenge, raw integer value data from the data acquisition is being simulated and must be converted to data in physical units. The new "digi" layer will hold the raw data and the old "hit" layer will continue to contain the post-unit-conversion data. Calibrations will be applied during this conversion.

There is in fact another layer below digi, for example the DFADC250 objects, which contain information direct from the data acquisition. The event recorder will be writing data at this lowest level to the output files. The translation tables relating geographic channel information to detector channel information are used in the conversion to the "digi" objects. In this scheme the raw data written from the experiment will not depend on the translation tables.

Spiral Track Reconstruction

Paul presented preliminary results on a new effort to reduce the number of ghost track produced by tracks which spiral in the CDC. See his wiki page for details.

He showed several events where the current tracking finds multiple tracks from a single generated track and where the new code finds one. He also showed average multiplicity from old and new code at the track candidate level and at the time-based track level. He also plotted the distribution for the track parameters at the track candidate level (all tracks were thrown with 150 MeV/c at 90&degrees;). The fraction of times where at least one "roughly close" track was found turns out to be similar between the two codes.

The results look very promising. Although these problem tracks are not a huge fraction of the phase space, for some reactions they may contribute significantly. There is a suspicion that they caused some of the crashes seen during the data challenge. One thing that remains to be studied is the effect of the new algorithm on multi-track events; we do not want it to be too good at eliminating tracks!

Paul has been consulting with Simon on this work. Simon will likely take over this work at some appropriate time in the future.

New HDDS Version

Simon has been working on adding the cooling manifolds on the FDC and fixing some issues with the BCAL geometry. These should not affect the nightly builds much and we are not ready to issue a new tagged release of the geometry.

Magnetic Field Switch

David intends to generate a new magnetic field map for 1350 A and the correct polarity. We will adopt it when it is ready.

Report from the Work Flow Meeting

Chip Watson has convened a committee to look at work flow tools and data provenance solutions for general use at the Lab. By this he means

  • Work Flow Tools: standard automation of large-scale, multiple job submission and tracking
  • Data Provenance: having information on how data was processed and archiving tools necessary to re-analyze data in the future.

Scientific Computing at JLab is willing to contribute resources to design and implementation of a system. The committee consists of representatives from each of the halls and from Scientific Computing. David and Mark are representing Hall D. The first meeting is next week, on Wednesday. Progress on the effort will be reported back to the collaboration as work goes forward.