ROOT Team meeting

32/1-A24 (CERN)



Show room on map
    • 11:00 AM 11:10 AM
      News 10m
      -SFT review 30 September, see agenda at As you can see from the presentations, the scope of the review is to address the readiness of the tools required to support the physics analysis of the LHC experiments, in particular Geant4 and ROOT, now that we are getting close to the start of data taking. The review should also address organisational and resources aspects and establish whether they are adequate for the long-term, continuous support to physics analysis of the LHC experiments. Impact from support to new projects, such as the Linear Collider Detector and other experiments, will also be analysed. ---------------------------------- -Bertrand will be away until October 12 (important in case you have problems with TRecorder) -------------------------------- -AForum: long discussions about releasing new versions of TMVA/RooFit with old versions of ROOT and vice-versa. Trying to understand why it takes so much time to 3 experiments to move to a new version of ROOT. Good contribution from CMS. --------------------------------- -LLVM: Philippe, Axel and Fons will participate to the LLVM users meeting end of September Axel has identified a potential candidate to work on the project as a doctoral student --------------------------------- -Technical student committee: we have requested Katerina Opocenska to work on the parallel file merging problem. Probably money not available before January. ----------------------------------- -other conferences -Varna: NEC2009 -Frascati: CFDT2009 : HEP and Fusion (ITER) communities. Rene will talk about I/O and data bases experience in HEP -Lorenzo will participate in the IEEE NSS conference with a presentation on ROOT statistical software ----------------------------------- -September 23: Volker Lindenstruth and Matthias Kretz visit. They have implemented a vector package optimum for parallel systems. Lorenzo will evaluate the product to be ready for the discussion. ---------------------------------- -Anar's week. to implement a better version of TMemStat (memory statistics) a lot of progress, ideas for the visualization. Note the Anar's proposal to move from SVN to GIT. For more details on GIT, see
    • 11:10 AM 11:30 AM
      ROOT 5.25/02 release status 20m
      Review of the status and developments planned before September 25
      General Comments =============== -Make sure yo are ready NOW with the release notes (and LIST OF CONTRIBUTORS) -if you had one or more patch in the 5.22 or 5.24 release patches, please make an ffort to improve your QA procedure. See: CINT: Axel =========== -status of editline; still portability problems -what else in CINT? CORE: Fons =========== -makefile/configure status (see also gviz) -test with latest gcc versions? -check make static. check if roota really works I/O & Trees: Philippe, Rene ============================ -still some work to do on the cache (TTree::Process and SetCacheEntryRange) -TFormula/TTreeFormula stability after Philippe's changes -May be! automatic activation of call to TTree::OptimizeBaskets -must understand cpu consomption on client of TWebFile (Fons) MATH: Lorenzo ============= -Any mods expected? -check with roofit/roostats. Lorenzo to check with Wouter asap -FitPanel recorder file - add improvements in building static representation of symmetric matrices - add new sparse bin data structure for fitting which combines the empty bins. Can be used for performing fitting efficiently multi-dimensional histograms or THNSparse. - various improvements in RooStats/RooFit to be ready for a tutorial sessions in mid-October GRAF2D : Olivier ================ -a long list of savannah reports. Which ones will be fixed -status of gviz and installation of graphviz on Linux when linking with libgvc.a must compile with -fPIC -recorder sessions GRAF3D : Matevz =============== -Any mods expected? -recorder session PROOF: Gerri ============= For the release, in addition to a couple of pending bug fixes, there are two developments that I would like to have in: - The extension of the 'multiuser' mode following the model used on the Grid; this feature is requested by Alice and it is almost ready. - The dataset validation in parallel (currently the dataset manager does it serially and several people complained that this may be long). These should be done next week. I may also have a recorder file for a simple PROOF-Lite session.
    • 11:30 AM 11:50 AM
      Recorder files for the release 20m
      -to be ready for the release -3 recorder files for graf2d (Olivier) -1 recorder file for Fitpanel (Lorenzo) -1 recorder file for graf3d (Matevz) -1 recorder file for the GUI in general (Bertrand) -1 recorder file for the GUI Builder -gviz/graphviz installation complete and well tested -TGraphStruct/Node:SavePrimitive implemented Ideally all recorder files should come with a test suite to verify that the graphics generated by the recorder is OK (a la stressGraphics). For example the FitPanel test suite should test that the objects generated and the fit results are identical to the original session.
    • 11:50 AM 12:00 PM
      Review of Savannah bug reports 10m
      -review of Savannah bug reports. It would be nice if the following list could be reduced by at least 50% before the meeting. statistics last Monday (left) and today (rigth) -couet : 23 23 -pcanal : 14 14 -moneta : 11 12 -rdm : 10 10 -axel : 10 10 -ganis : 7 6
    • 12:00 PM 12:10 PM
      ROOT/PROOF to be included in EGEE RESPECT 10m
      Dear Rene Brun, I'm contacting you to inform you that the ROOT/PROOF toolkit has been proposed to be included in the EGEE RESPECT program. If you are not familiar with RESPECT (Recommended External Software for EGEE CommuniTies) you can find more information at EGEE project's web site: In brief RESPECT is a list of software tools that are being developed independently of EGEE but taking advantage or somehow extending the Grid services offered by gLite software and the EGEE grid infrastructure. Inclusion in RESPECT means that EGEE endorses these tools by recognizing their utility for the broad grid user community. The proposal to include ROOT/PROOF was submitted by Torsten Rathmann from DKRZ, who is currently acting as a Regional Coordinator of the EGEE NA4 activity in Germany & Switzerland. Torsten has already provided all the information needed in order to consider ROOT/PRROF for RESPECT but since he is not directly affiliated with the development team we have to confirm that you agree to include the toolkit in RESPECT. From your part this means a commitment to continue supporting the software and provide any required feedback to user problems (which I suppose you already do). Your contact details and the suggestion to get in touch with you for this purpose was give to my by Patricia Mendez Lorenzo. Looking forward for your comments and response. Kind regards, Vangelis Floros
    • 12:10 PM 12:30 PM
      AOB 20m
      Mini workshop about graphics (Olivier)