LHCb Core Software Meeting

Date and Location

10:30 - 11:20
CERN (2-R-030)


Gregoire Astruc, Illya Shapoval, Marco Cattaneo, Marco Clemencic (minutes), Markus Frank, Martin Plajner, Rob Lambert, Wojciech Fijołek



  • Wojciech Fijołek joined us as Technical Student
    • he will be working on the migration of the LHCb web sites to Drupal
  • From the LIM
    • requested release of LCGCMT 61, which should be ready around the end of next week
      • it should contain a newer version of ipython and other Python modules (to be tested from the nightlies)
      • there should be a build on gcc 4.5 so that we can start testing it for a later migration
    • the nightly build slot dev1 (LCG 59) is going to be moved to LCGCMT 61 (we will have to bind one of our nightly builds to is as soon as it is ready)

Round Table

Marco Cl.

  • Gaudi: preparing for the release of v22r3
    • several minor or backward compatible patches in the trunk to be merged into the v22 branch
    • the new parsers gave more troubles than hoped for
      • backward/forward incompatible: old LHCb doesn't work with the new Gaudi, and an adapted LHCb doesn't work with the old one
        it can be fixed re-introducing in Gaudi the headers that have been removed (adding a deprecation warning) and modifying the client code using #ifdef branching
      • there is a reduction in the size of the optimized libraries in Gaudi (~5%), but the debug version are two times bigger; that's probably due to the parsers for STL classes that were not used before (like list, set etc.) or to the need of splitting the .cpp file of the parser in several small pieces to improve compilation time
        we agreed to remove the unnecessary parsers
  • Testing profile-driven optimization with gcc (task #21707)
    • it's clear that we can gain from it, but there are several other tests to be performed before being able to use it in production.
      Marco Ca. believes (given the technical obstacles) that it is more urgent to fix the code first (avoid pointless decoding etc.), and it is important to verify that the gain observed is valid for other strippings and repeat the exercise for the other applications.
      Rob would like to push further this optimization because it should be somehow orthogonal to the fixes in the code and easier to be used in production (the fixes require touching several files, while this optimization is centralized). Marco Cl. feels the need of a well defined set of "Reference Workloads" to support any attempt to improve performances.


  • When is Gaudi supposed to be released?
    About the 15th of August. If It happens that Gaudi is late and a release of Online is needed, he will build it on the old Gaudi.


  • Problem with CondDB access to IN2P3 solved (the password used when the database was restored was the wrong one).
  • IT is planning to upgrade the Oracle installations to the version 11g (January 2012), so Illya have been contacted for help in the validation tests.


  • Prepared some new tools to be included in LbScripts (bug #83487). They need to be validated by Hubert (Marco Cl. will have a look while Hubert is away).

-- MarcoClemencic - 03-Aug-2011

Edit | Attach | Watch | Print version | History: r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r1 - 2011-08-03 - MarcoClemencic
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LHCb All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2021 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback