GlueX Offline Meeting, March 20, 2013

From GlueXWiki
Jump to: navigation, search

GlueX Offline Software Meeting
Wednesday, March 20, 2013
1:30 pm EDT
JLab: CEBAF Center F326/327

Agenda

  1. Announcements
  2. Review of minutes from the last meeting: all
  3. Making REST skims: Paul M.
  4. Simple Email Lists: Mark
  5. No more hardwired run numbers: David
  6. Switch back to use of public version of ROOT at JLab?: Mark
  7. HDDM changes: Richard
  8. Action Item Review
  9. Review of recent repository activity: all

Communication Information

Video Conferencing

ReadyTalk Desktop Sharing

You will need a web browser open to view the computer desktop in the meeting room at JLab.

  1. Go to http://esnet.readytalk.com
  2. In the "join a meeting" box enter the Hall D code: 3421244
  3. Fill in the participant registration form.

Telephone

To connect by telephone:

  1. dial:
  2. enter access code followed by the # sign: 3421244#

Slides

Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2013-1Q on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2013-1Q/ .

Minutes

Present:

  • CMU: Will Levine, Paul Mattione
  • IU: Kei Moriya, Matt Shepherd
  • JLab: David Lawrence, Mark Ito (chair), Simon Taylor, Elliott Wolin, Beni Zihlmann
  • UConn: Alex Barnes, Richard Jones

Making REST skims

Paul reviewed the change that he checked in to allow users to skims of REST format data. Previously the REST-writing capability was bound up with other code used during the data challenge. See his email for details.

Simple Email Lists: Mark

Mark reviewed the mechanism we are using for three notification-type email lists:

  1. nightly builds of GlueX software
  2. b1pi reconstruction tests
  3. single-track reconstruction tests

The system is described on his wiki page. The idea is to have a light-weight system suited for cron jobs and targeted at a limited audience.

No more hardwired run numbers

David reviewed his announcement of changes that removed a few instances of hard-wired run numbers in our code. Up until recently there was no run dependence in calibration constant look-up, so there was little practical effect.

Switch back to use of public version of ROOT at JLab?

Mark proposed that at JLab, we go back to using the IT-Division-provided versions of ROOT. Some months ago we switched to using privately built (Hall D specific) versions. The intent is to support any effort IT Division makes to support physics analysis in general by using the products they offer, i. e. demonstrate that we can play nice with others. There was general support for this move. Mark will work on making the change.

HDDM changes: Richard

We reviewed some proposed changes from Richard to rationalize the data model for Monte Carlo data to more clearly identify information that is known only from the simulation and would not be available in real data. This information had been incorporated in a somewhat ad hoc fashion. The need for this information is real; this is an attempt to clean up the implementation. See his talk, presented in various versions at a series of meetings held over the past couple of weeks, for details of the proposal. Feedback from David, Beni, Simon, Paul, and Mark has been incorporated in the plan.

Richard wants to do these changes before doing a conversion of the simulation from Geant3 to 4. He hopes to have this set of changes done by the Summer.

JANA 0.6.6

David announced that he has tagged a new version of JANA, version 0.6.6. Simon and Paul have been using it. New features include:

  • resources
  • start-up time improvement for multi-threaded jobs
  • warnings due to conflicting autoconf makefiles eliminated

Computer Resource Request

Graham Heyes has asked us to estimate the resource use at JLab for the FY14, 15, and 16. From his message:

  1. What computing capacity do you require in units of cores of the current generation of machines in the Farm.
  2. What disk capacity do you need, volatile, work etc, units of terabytes.
  3. How much data do you anticipate writing to tape, units of terabytes.

Mark will work on drafting a response. Input from the collaboration welcome.

Policy on Virtual Machines at JLab

David mentioned that Andy Kowalski of CNI recently circulated a draft proposal for use of virtual machines at the Lab. Previously, these have been forbidden and he is outlining situations where their use may be allowed. The policy is aimed at maintaining cyber security by insuring that only vetted operating systems are running on Lab-owned equipment. There is some DOE-oversight context to these policies as well. There has been a lot of internal discussion at the Lab about the document since the use of virtual machines is...well...useful in many contexts and the policy would not allow users free rein to install them. Such use is not unprecedented, despite the existing policy.