28th March 2025 Minutes

- Intro (Alberto): No questions
- Kernel (Gabriele): 
    - Marilena: On slide 4: is the voxelization fix solving an issue present in 11.3 only or also in previous releases?
        - It is only a patch to 11.3, and it affects the voxelization of the geometry between multiple runs
- EM physics (Vladimir): no questions
- Hadronic physics (Alberto): 
    - Marilena: Questions on the Birks law tuning issue reported by ATLAS. Do you have already ideas on how to tackle this issue? 
        - Not yet, ideally one could find an energy point in between the 2 models and tune at that energy. Otherwise we should change the Birks law threatment and make it dependent on the energy/particle type, but that would be too complex.
- R&D program (Witek): no questions
- User requirements (Krzysztof)
    - General comment on the last slide, on the number of solved issues vs the still open one (or without person power). The Geant4 collaboration is reduced in person power and it is struggling to fulfill the long list of users requirements:  It's important to open to the users community and invite them to contribute to Geant4. How to contribute: https://geant4.web.cern.ch/contributing/
* Behavioural change of the FTFP_BERT_HP physics list between Geant4 v10.6 and 11.1 in the EM simulation (Doris Kim)
    * Alberto: It is very likely and hadronic problem the one you are observing. Do you integrate regardless on the time or do you have a time window in which you record? 
        * Integrate
    * Alberto: then one possible explanation of what we see here is that between 11.0 and 11.1 (to be checked) a time cut was introduced on the radioactive decay.  In 10.6 you would get the contribution of any kind of nuclear fragments.  This cut is very high (double of the age of the universe). So the effect would be small, but visible in some cases.  If this is the issue you would get a more physical behaviour in the new G4 version, because in reality your time window will be very small. Alberto will send Doris a command on how to disable this cut. The experiment should have a time window that is close to reality
        * Doris: it is possible that there's a hidden window in the reco tool, but will investigate to see if this is shown in this study
    * Alberto: you showed an impressive statistics of data. Is the MC statistics compatible to data? Is the ration 1:1 as in LHC experiment?
        * Doris: for physics is about 1:1
    * Alberto: do you have problems in generating such high statistics?
        * Doris: so far it's ok, but if the luminosity will be increased they might have troubles. For now we are running at lower luminosity. To use Fast Simulation we would need to optimize optical photon simulation for Cherenkov
    * Krzysztof: it would be maybe beneficial to use 11.1.2 instead of 11.1.1 (so the latest patch available for each release)
        * Doris: Sure, will ask our librarian
* GeoModel: an experiment agnostic tool suite for detector description (Marilena Bandieramonte)
    * Gabriele: very nice tool. I saw that it requires QT5, is there any plan to update to QT6?
        * GeoModel supports also QT6 and work is ongoing to deprecate QT5
    * Alberto: it looks a nice and interesting tool. It look like a competitor of DD4Hep, what are the main pros/cons of GeoModel w.r.t. DD4HEP?
        * Marilena: cannot comment much on DD4HEP because I don't know it very much. What I can say is that years ago ATLAS has evaluated moving to DD4Hep and has decided not to move to it. Main reason being that GeoModel has been in usage since more than 20 years and all the features are already implemented and fulfilling all the experiment needs. It's a solid development. Also it is modular, with minimal dependencies, and portable. 
    * Krzysztof: the speed of the visualization was impressive. Can you say how many volumes are shown? 
        * Marilena: this is the whole ATLAS geometry, so many volumes. Don't know exactly how many, but of the order of millions. Another nice feature of GeoModel is the fact that it can be installed in your laptop and MacOS/Linux/Ubuntu are supported and tested. It can be also installed on Windows but it is not tested extensively. This is important to cut the development time. 
    * Krzysztof: Can you also do simulation in this framework, are you able to display tracks?
        * Marilena: Yes that is possible, we provide 2 plugins, one for tracks and one for hits. Those write out hits on HDF5 files that can be uploaded to gmex and visualized there in a static way (not like Geant4)
    * Krzysztof: What about SD?
        * Marilena: FullSimLight can be extended with SD plugins, but we haven't implemented any specific one for ATLAS. We are extending the scoring functionality to write out histograms per subdetectors-regions
    * Alberto: If I am not wrong GeoModel was presented in one of the FCC meetings, what was the reaction there?
        * Marilena: It was presented by another colleague. Short time for the presentation, they didn't get back to us yet. 

OAB from John Apostolakis: the next Geant4 beginners course May, it will be linked to the Geant4 website. Please spread the word to people that might be interested.

Next G4 technical forum October 2025, during the G4 Collaboration meeting at JLAB