LHCb Core Software Meeting

Date and Location

April, 30 2008
10:30 - 12:00
CERN (2-R-030)


Gloria, Hubert, Joel, Marco Ca., Marco Cl. (minutes), Markus, Philippe, Thomas


Patrick (phone)



  • LCGCMT 54e and 54f have been released (Gaudi v19r9 is base on 54e)
  • A software stack without SEAL is available in the nightlies (slot dev). Marco Ca.: It is the one we have to use for next Gaudi.
  • Marco Ca.: In the Architects Forum it has been agreed that the next version of ROOT will drop support for files produced with ROOT version < 5.
    In LHCb, only DC04 data has been produced with ROOT < 5 (3).
  • Joel: During the last linux certification meeting, even if the experiments didn't ask for certification of SLC5, it has been decided that there will be a certification and SLC4 will be phased out during 2009.Q1.

Software Releases

Gaudi (Marco Cl.)

  • We will have the latest feature-release of Gaudi in the v19 series around the end of May (v19r10). Except for bug-fixes, it will be the version used for first data. After that we will start the v20 development (that should be able to read the data produced with v19, for testing).
  • To be able to build Gaudi against the slot dev of the LCG nightlies, we need to apply the patch by Markus, but it breaks the build on LCG 54. To solve it, we will have a separate tag for LCG 54 in the nightlies.

Marco Ca.: We need to start testing the new ROOT schema evolution in Gaudi v20.

Gloria: Philippe Canal is preparing a version of ROOT that allows transparent conversion between HepMC1 and HepMC2. We have to push to get it in a patch release of ROOT (to be able to pick it up in Gaudi v19), otherwise we will not be able to switch to HepMC2 before the end of the year.

LHCb, Boole, Brunel (Marco Ca.)

  • New release last week (for CCRC)
    • new STDev does not work yet (problems with new geometry)
  • Next release will feature the new Geometry.

Marco Cl.: Adlene Hicheur asked for a the Magnet conditions. It is not yet possible to have it inside the DB (Marco Cl. needs to talk to Carmen first), but the idea is to have the "requested" and "read" current. The MagnaticFieldSvc will be modified to use them. Gloria suggested the implementation of a small IMagneticFieldSvc dispatcher to handle to two maps we have. It has been agreed to have the name of the map in the CondDB too.

Gauss (Gloria)

  • Releasing new DEC fiels.
  • New release of Gauss on the latest LHCb will be released next week (on a patched version of Geant4), for testing. Then Gloria will start with the new Geant4.
  • Patrick Robbe prepared some tests with scripts comparing histograms with reference histograms. We agreed to have the reference histograms in CVS, and the tests integrated with the QMTest infrastructure.

DaVinci (Marco Ca. for Juan)

  • DaVinci v19r12 released yesterday.
  • The version in DEV contains the changes in the API of the PhysDesktop discussed last week.

OnLine (Markus)

  • New version being built.

Panoramix (Thomas)

  • Panoramix v15r13 is ready.
    • Add-ons for Online (e.g. simplified menus).

Round Table

Marco Cl.
Working on the database of releases, discussed some time ago, to have proper handling of runtime dependencies.
The DB is based on an XML file and a python module that reads it and the queries are done in memory. The database can also handle dependencies between projects, and it is planned to be used by install_project.py too.
Marco Ca. suggested to have a web page describing the releases. Marco Cl. says that it can be dynamically generated from the XML file.
  • Current software stack almost completed.
  • Problem with Doxygen. Hubert is preparing a pattern to be able to produce Doxygen documentation for single packages for tests.
  • Software installation (install_project.py):
    • using tar command instead of tarfile module if Python 2.4.4 (broken module)
    • LHCbGrid project disabled on SLC3 because of missing gfal
    • reduced the size of project tarballs by excluding .o and .a files
  • LHCb Nightlies are getting ready
    • few minor problems to be sorted out
    • we need some space on AFS to be able to copy the output of the nightlies
    • we have to define the content of the slots

-- MarcoClemencic - 30 Apr 2008

This topic: LHCb > WebHome > LHCbComputing > CoreSoftwareMeetingMinutes > CoreSoftwareMeetingMinutes20080430
Topic revision: r1 - 2008-04-30 - MarcoClemencic
This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback