CLD with ARC

[atolosadelgado]

BEGINRELEASENOTES

• CLD_o3_v05 is created. This option evolves from option 2 version 05. The main characteristics with respect the mother version are:
  • ARC subdetectors included between the trackers and the ECal.
  • Trackers are shrunk by 20 cm in radial and Z direction to accommodate the ARC. The space between layers is reduced proportionally, and the angular coverage is preserved.

ENDRELEASENOTES

Hi,

components from CLD (o2, v05) were taken to build the option 3, which contains the ARC detector, between the tracker and the ECAL. Modifications to the size of the outer tracker barrel, outer tracker endcap and inner tracker endcap were detailed in this presentation.

Overlap check with 10k points passed (it took 4h).

I am doing now some efficiency plots, but I think it was a good moment to open the PR.

[atolosadelgado]

I just noticed that the volume names, constant names, and the readout name for the ARC do not match the style of the CLD.

I will change it right away, I only need some minutes to do it.

[atolosadelgado]

Changes in the naming are done, tests passed

[atolosadelgado]

Should some release note comments be filled in the OP?

[jmcarcell]

Can you fix the conflict @atolosadelgado? It's a very simple one line in the README

[atolosadelgado]

I still do not understand why I get different simulation output if I use LCG, Key4hep or my local DD4hep, but since it is surely unrelated to this repository, please feel free to merge this PR at your earlier convenience :)

[atolosadelgado]

Please, do not merge, I have just found overlap of sensor in cell 11 of the barrel... sorry for the noise

[jmcarcell]

@atolosadelgado if you have Alma 9 the latest nightlies have Geant4 without Vecgeom. CentOS 7 will have to wait a bit and all the workflows here are CentOS 7...

[atolosadelgado]

thank you very much! I just tested the overlaps in lxplus9 and now the test passes :):)

This is the code to reproduce the test,

source /cvmfs/sw-nightlies.hsf.org/key4hep/setup.sh
cd /tmp/
git clone -b CLD_with_ARC [email protected]:atolosadelgado/k4geo.git
cd k4geo/
cmake -B build -S . -D CMAKE_INSTALL_PREFIX=install
cmake --build build -j 4 -- install
export LD_LIBRARY_PATH=$PWD/install/lib:$LD_LIBRARY_PATH
cd build/
ctest -V -R t_test_ARC_o1_v01_overlap

At the beginning I though the problem was the unit system and I added a dummy offset in sensor position to try to fix the overlap problem. It turns out that the problem is VecGeom and that offset is not needed. I have removed that ad-hoc offset in the most recent commit. I believe this PR can be merged now.

[andresailer]

Now there is a test failure with the overlap check

[jmcarcell]

LCG_100 has Geant4 with Vecgeom enabled, which could explain the issues. The CI test with the nightlies just passed (CentOS 7 - Geant4 without Vecgeom since a while ago).

[atolosadelgado]

Thanks Juan for the clarification. I have run the overlap test in centos 7 (LCG latest view dev4) and alma9 (LCG and key4hep) and the test passes.
LCG 103 and before use Geant4 with VecGeom, so the overlap check will not pass.

[andresailer]

Ok, we should update the stacks used for the tests then.

[jmcarcell]

Here or can we merge this PR @andresailer ? k4geo is one of the few missing tags

[andresailer]

Yes, let's merge it and update the CI bases after