CoreSoftwareMeetingMinutes20090429 < LHCb

LHCb Web>LHCbComputing>CoreSoftwareMeetingMinutes>CoreSoftwareMeetingMinutes20090429 (2010-06-11, PeterJones) (raw view)

EditAttachPDF

<noautolink>
---+!! LHCb Core Software Meeting
---++!! Date and Location
[[http://indico.cern.ch/conferenceDisplay.py?confId=55752][29 Apr 2009]]%BR%
10:30 - 12:00 %BR%
CERN (2-R-030) %BR%
---++!! Attendees
Gloria,
Joel,
Juan,
Karol,
Marco Ca.,
Marco Cl. (minutes),
Markus,
Olivier,
Patrick (EVO),
Philippe,
Thomas
<!--
---++!! Apologies
-->
---++!! Subjects
%TOC%
---+ News
   * LCGCMT 56a being prepared%BR%
      We should try to get fixed the warnings produced by gcc 4.3 (currently ignored in the nightly builds)

---+ Software Releases
---++ Gaudi (Marco Cl.)
   * Marco Ca. is building Gaudi v21
      * Some problems with the scripts to make the release (now fixed)
      * Problems with Win32:
         * Usual minor compilation problems (fixed, but we need the nightlies)
         * LbLogin does not work out of the box on Windows (missing env. variables, etc.)
         * Being rebuilt on the terminal server (no FORTRAN available on the machine Marco Ca. was using... do we drop HbookCnv?)
   * Written a small script to try to replace the "build dependencies" in CMT: speed up from ~2min (v1r20p200702) to 3.5s (the most recent version is faster, 6s, but doesn't look outside the project). The algorithm will be sent to Grigori for discussion.

---++ LHCb, Boole, Brunel (Marco Ca.)
   * Will release as soon as Gaudi is released
      * lhcb1 will contain the pre-release for MC09
      * lhcb2 will contain the upcoming v27 version of LHCb

---++ Gauss (Gloria)
   * Cleaning up
   * Working to move to the next LHCb version (keeping an eye on the MC09 version).

---++ HLT (Patrick)
   * There has been some discussion about the build-time dependencies between projects due to the Python configurables
      * The organization of configurables needs a review.

---++ Online (Markus)
   * Ported ONLINE to gcc-4.3 and SLC5
      * SLC5 is missing the X11 (Motif) development packages, requested to linux.support
   * Discussed with Axel the issue about the warning produced in the dictionaries (=casting between pointer-to-function and pointer-to-object=). The problems has been fixed in the HEAD version of ROOT.%BR%
      Marco Ca.: we should get the fix in the patch version too.
   * Problems trying to set up the multi-process integration tests, QMTest doesn't seem to be good enough. There is also a problem if a tests is run many times in parallel.%BR%
      Marco Cl.: the GaudiTest wrapper class may not be suitable to the specific use case, but a dedicated wrapper can be written to cover the specific needs.

---++ Panoramix (Thomas)
   * Guy produced a new OpenScientist. Panoramix gets some compilation errors.

Marco Ca.: Some tests are failing. Thomas will look into them ASAP.

---+ Round Table
   $ *Karol*: The display of uncompleted platforms is not yet automated (currently triggered by hand).
   $ *Joel*: 
      * Releasing and installing DIRAC v4r12.
      * Started MC09 test production, but with problems.
      * Checking to get access to some OpenLab machines for the profiling of out software.
   $ *Philippe*: 
      * Meeting next week to discuss the CORAL/LFC problem.
      * We need to settle the pending issues about the usage of LHCbGrid. We need to include the cross-OS compatibility libraries in it.
      * We need to set up a team to make our applications produce a summary to be able to easily validate the execution (instead of parsing the standard output of the job).
   $ *Olivier*: Still working on the improvement of the start up of Brunel: reduced (the detector description part) from 35s to 19s. Main changes are
      * Moving from SQLite to XML files
      * Smarter caching in the XML parser
      * Improve XML: merge files, avoid the usage of entities for macro expansion

Marco Ca.: How do we proceed?%BR%
Some code could be committed, but it is better to have a more final picture of the actual gain. One thing that could be already done is to prepare the snapshots for the HLT in form of XML files instead of SQLite files (it may be tricky to manage the DDDB and LHCBCOND independent tags since the XML files need to be in the same tree). Eric should be included in the loop, since he is in charge of making those snapshots.
----
</noautolink>

-- Main.MarcoClemencic - 29 Apr 2009

Topic revision: r3 - 2010-06-11 - PeterJones

LHCb

Public webs

Welcome Guest

- Cern Search
- TWiki Search
- Google Search
LHCb All webs

Copyright &© 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