DD4hep Developers Meeting

Thursday 21 Jul 2022, 15:30 → 16:00 Europe/Zurich

zoom

Andre Sailer (CERN), Markus Frank (CERN)

Please Join DD4hep developer's meeting via Zoom

Present:  S. Banerjee, E.Eren, M. Frank, J.Smiesko, A. Sailer, T.Madlener

1) Pending problems

1.1) Pending problems reported by Sang Hyun Ko:
     - Cannot setup Birk's constant as material property.
     - No analogy to G4PVParametrized. Unclear if a solution is possible.
1.2) PluginFactoryBase::make_return problem with longs (Issue 875)
1.3) ROOT issue: for G4 units the radiation/interaction length is wrong by a factor 10
       -- For CMS this problem does not show up.
       -- Should only enter AFTER simulation (Geant4 materials do not have the radiation length as input)
       -- Need to address this issue early next year.
1.4) DDDigi TBB code needs migration to new oneAPI TBB interfaces (Issue 901):
       Clash between Intel oneapi TBB and classic TBB.
       -- Wouter promised to look into.
1.5) Issue 913: XML parsing with non-standard linux-locale.
1.6) Issue 918: Should be addressed by PR920. Waiting for feedback.


2)    Closed issues and fixed problems
----

2.1) PR928: CI builds for MAC removed.
       MacOS on github does no longer allow to mount cvmfs with fuse.
2.2) PR929: Improved parsing for component properties
2.3) PR931: Variant parameters
2.4) PR932: Fix typos in user manual
2.5) Issue 909: ddsim gets stuck infinitely on some events (reported by T. Madlehner).
       Issue understood. Had to do with PDG particle properties.
       Proposal is to implement a sanity checker to pre-check such occurrencies.
2.6) PR933: GFlash and fast simulation handling in DDG4
        Will need some testing though. The question of parallel worlds will have to solved as well.

3) Round table
----
3.1)  Markus:     ---
3.2)  Andre:       Worked on PR920
3.3)  Sunanda:   No issues. C++ with optimization O3 gives different result.
                          Most likely not connected to DD4hep.
3.4)  Engin:       Plans with GFlash. Need to see how it will work with machine learning techniques
                          and a realistic detector scenario. Wants to compare machine learned shower simulations
                          with GFlash and detailed simulation.
3.4)  Jurai:         ---
3.5)  Thomas:    ---


4) AOB
----
Andre will create a new tag 1.21 for the ACTS folks.
 

Next meeting:   04. August 2022
 

15:30 → 15:50 Around the table