GlueX Offline Meeting, December 10, 2014
GlueX Offline Software Meeting
Wednesday, December 10, 2014
1:30 pm EST
JLab: CEBAF Center F326/327
- Work Disk Usage Report
- GlueX Project Accounts for JLab Farm: username = gxproj1, gxproj2, gxproj3
- Discontinue SeeVogh Contract(?)
- Software Review - February 10+11
- Review of minutes from November 12 (all)
- track, itrack, and the Truth (Richard, Justin)
- Data Challenge 3
- Commissioning Simulations
- Offline Data Monitoring (Paul M.)
- Version Management II (Mark)
- Action Item Review
- The BlueJeans meeting number is 968 592 007 .
- Join the Meeting via BlueJeans
Talks can be deposited in the directory
/group/halld/www/halldweb/html/talks/2014-4Q on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2014-4Q/ .
- CMU: Paul Mattione, Curtis Meyer
- FIU: Eric Pooser
- FSU: Aristeidis Tsaris
- JLab: Alex Barnes, Mark Dalton, Mark Ito (chair), Manuel Lara, Will McGinley, Kei Moriya, Nathan Sparks, Mike Staib, Simon Taylor, Beni Zihlmann
- IU: Matt Shepherd
- MIT: Justin Stevens
- NU: Sean Dobbs
- Work Disk Usage Report After we ran short of work disk space a month or two ago, Mark I. started work on a web page to report on large files, large directories, and users with large disk usage. It is now updated nightly. Reports on other disk systems are planned.
- GlueX Project Accounts for JLab Farm Mark I. has asked the Computer Center to set up three new accounts with elevated priority on the farm for GlueX projects. The usernames are gxproj1, gxproj2, and gxproj3. The solves the problem we have been having where multiple projects running from the "gluex" account were getting in each others way (long queue waits for the second one in). Kei is running the offline monitoring jobs from gxproj1 right now.
- Discontinue SeeVogh Use Mark I. told the Computer Center that GlueX is not interested in renewing the license.
- Software Review Rolf Ent is organizing the next Software Review for February 10 and 11. Curtis commented that we have a lot of things to show now that we are analyzing real data from the current commissioning run.
track, itrack, and the Truth
Justin told us that the changes that Richard introduced to give sensible truth information for showers in the BCAL are working well.
Data Challenge 3
We decided on a schedule for Data Challenge 3.
- Before Holiday Shutdown: start generating simulated raw data and stage it on the Tape Library. Try to get a lot of this done over the shutdown.
- After Holiday Shutdown: start setting up the reconstruction jobs to work on the data.
We discussed trying to use real data for this challenge now that we have some. We decided against it; the real data presents multiple challenges of its own that are better dealt with in a different context.
We reviewed the goals:
- Handle data from the tape library
- Multi-core reconstruction
- Use EVIO-formatted data
Sean noted that the geometry in the Hall has changed again recently. We need to make the updates and re-run simulations with the new conditions. Changes include more shielding around the beamline and relocation of the active target downstream of GlueX.
Offline Data Monitoring
Paul took us through a tour of the offline monitoring plots showing new histograms, anomalies, and other points of interest.
Mark I. thought that a similar presentation would be nice at the morning run coordination meeting. That may happen.
Run and Data Conditions
Sean presented some slides on the story behind the recently released run conditions web page. He described the origin of the various fields and also listed other meta-data (such as software versions) used in the offline monitoring jobs.
Some suggestions for improvements were offered. In that context, Justin asked about the Run Conditions Database that Dmitry Romanov was working on. We have heard that it contains similar information, but no one at the meeting has seen it. Mark reported that he talked to Dmitry about this recently and agreed to get a version deployed somewhere so that people can have a look so we can decide where to put future effort.
Version Management System
Mark presented slides on a system for specifying a compatible set of versions of the various software packages we use for GlueX data. An XML file can be used both to direct a build of the software and to generate the environment to use such a build. The system has been tested (but not exhaustively) and is already used by the gluex_install scripts.
- Update the Hall geometry. (Simon)
- Run more commissioning simulation (Sean, Mark I.)
- Deploy Dmitry's run conditions database so people can have a look. (Mark I.)
- Finish roll-out of the version management system. (Mark I.)