GeoModel 4.5.0 changes ATLAS full simulation output

This is a follow-up to the discussion we had in the ATLAS DD technical weekly meeting on August 28. It turns out that GeoModel 4.5.0 changed the output of ATLAS full simulation for Run2 and Run3 configurations. This was discovered when we attempted to switch Athena to GeoModel 4.5.0 in this MR. AthSimulation tests failed because of the GeoModel version update. Here is the list of the failed tests:

• CITest_SimulationRun2FullSimChecks_ctest
• CITest_SimulationRun3FullSimChecks_ctest
• CITest_SimulationRun2FullSimLegacy_ctest
• CITest_SimulationRun3FullSimLegacy_ctest
• CITest_SimulationRun3HitsMergeWithSort_ctest

I tracked down the source of these differences to !195 (merged) which changed the way volumes get calculated for several GeoModel shapes.

@evc, could you please comment on whether the changes in simulation output were expected? If you think this can be useful for you, then you can find reference outputs as well as the outputs produced with !195 (merged) at the locations listed below:

1. SimulationRun2FullSimChecks
   Reference: /cvmfs/atlas-nightlies.cern.ch/repo/data/data-art/WorkflowReferences/main/s4005/v1/myHITS.pool.root
   Changed: /afs/cern.ch/user/t/tsulaia/workspace/public/GeoModel-4.5.0/s4005/myHITS.pool.root

2. SimulationRun3FullSimChecks
   Reference: /cvmfs/atlas-nightlies.cern.ch/repo/data/data-art/WorkflowReferences/main/s4006/v1/myHITS.pool.root
   Changed: /afs/cern.ch/user/t/tsulaia/workspace/public/GeoModel-4.5.0/s4006/myHITS.pool.root

3. SimulationRun3HitsMergeWithSort
   Reference: /cvmfs/atlas-nightlies.cern.ch/repo/data/data-art/WorkflowReferences/main/s4007/v1/myHITS.pool.root
   Changed: /afs/cern.ch/user/t/tsulaia/workspace/public/GeoModel-4.5.0/s4007/myHITS.pool.root

Thanks.