LHCb Core Software Meeting

Date and Location

14 Jan 2009
10:30 - 12:30
CERN (2-R-030)

Attendees

Gloria, Hubert, Juan (EVO), Karol, Marco Ca., Marco Cl. (minutes), Markus, Olivier, Stuart, Thomas, Wouter

Apologies

Patrick

Subjects

News

From Architect's Forum:
  • Last officially supported platform of CERNLIB is SLC4: it will not be built for SLC5.
  • Deployment of SLC5: we have problems of communication
    • FIO talks to lxcert and not to AF (the complaints about SLC5 should have gone through lxcert)
    • New machines will have SLC5 installed (SLC4 cannot be used because it is too old: no drivers)
    • We need to check if we can run SLC4 binaries on SLC5

Specific Topics

How to run calibration on express stream? (Wouter)

The express stream (5Hz of events) will need to be processed in real time. Assuming 1s per event, we need to be able to process 5 events in parallel. The main problem is that the results of the processing must be combined to get the final result.

One possible way is to use the Python processing module recently being tested, but requires some work to set up the needed script. Currently Adlene is investigating this approach.

A more interesting way to go seems to be the File Summary Record. This approach requires some work too, but it is simpler: just prepare the DataObject specialization to be stored in the File Summary Record. Wouter will work on it.

Software Releases

Gaudi (Marco Cl.)

  • Gaudi v20r4 has been released.
Hubert
the check-out has been done using SVN.
  • Started the work on v21

Marco Ca. asked the time scale for a first version, but the work is in such an early stage that it is not possible to give estimates yet.

LHCb, Boole, Brunel (Marco Ca.)

  • LHCb, Rec, Lbcom are being released.
  • Boole and Brunel are almost ready (only some last checks pending). These are the version for production, only bug-fix will be allowed.
    • Not using the new Particle property Service to avoid the risk of conflicts.

DaVinci (Juan)

  • The tags should be ready by Friday. Hubert will make the release on Monday.

Panoramix (Thomas)

  • Tagged. It features mainly changes in Python scripts for Configurables and the new Particle Property Service-

Round Table

Marco Cl.
Fixed the long-standing problem with VeloAlignCond.
Hubert
  • Clean up in the tar-balls of LCGCMT, the size has been reduced to 1/2.
  • LHCb, Rec, Lbcom, Geant4 and Gauss are being built.
  • Changes in LbScripts
    • removed the extension of regular scripts (shell scripts to be sourced kept it)
    • aliases sourcing shell scripts locate them in the PATH (instead of hard-coded location)
    • python scripts used in shell wrappers are now in PYTHONPATH (started with "python -m")
  • Modified the getpack in $LHCBPYTHON to use ssh (ext) instead of kserver. The AFS release area will be modified on the 31st.
Stuart
  • 15k jobs running at the same time in Dirac3
  • The NFS misconfiguration problem is hitting us more and more frequently. The next version of LHCb have the possibility of switching on easily the local copy of the SQLite files (not enabled by default).
Marco Ca.
  • Chris Jones built and tested successfully the applications on OSX.
  • The Muon survey have been switched off in the latest CondDB tag because it has problems (Alessia is investigating).
  • Paolo Calafiura asked what we did to optimize our applications (for a talk he has to give at CHEP).
  • Sent a mail on Friday summarizing a coffee discussion (with Gloria, Marco Cl. and Thomas) about which tags to use for the processing. The information should come from the bookkeeping, but we have a problem with the definition of compatible tags (what does that mean?).
Wouter
Alignment production to be re-started (with Velo closed). Marco Cl.: the next version of SQLDDDB (v5r2) will have the global tags to be used for the production.
Olivier
Modified the UpdateManagerSvc to speed up the look up of already inserted conditions, gaining a 25% improvement in initialization time of Brunel.
Markus
Again problems with dependencies: DAQEvent depends, for a test, on DetDesc. DAQEvent should only contain RawBanck and (may be) ODIN, but it also contains some components. Marco Ca. will create a small project to contain the part of LHCb that is (strictly) needed by Online, to allow the mixing of different version of Online and LHCb. He will clean up DAQEvent too.


-- MarcoClemencic - 19 Jan 2009

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r2 - 2009-01-19 - 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 & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback