MTD geometry: fix BTL numbering scheme for scenario v3, input by Geant4, move unit tests to D110 scenario

[fabiocos]

PR description:

Fix of BTL numbering scheme for the latest geometry scenario v3, when DD4hep input is passed by Geant4. Addresses the failure reported in #45175 (comment) .

The whole set of unit tests for MTD geometry is updated to scenario I17 / D110, a corresponding update of the cms-data reference cms-data/Geometry-TestReference#14 is required.

PR validation:

Code checked both with standalone test code (not failing anyway even before) and test workflows 24834.0, 24834.911, 29634.0, 29634.911 .

[cmsbuild]

cms-bot internal usage

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-45211/40561

• This PR adds an extra 16KB to repository

[cmsbuild]

A new Pull Request was created by @fabiocos for master.

It involves the following packages:

• Geometry/MTDCommonData (upgrade, geometry)

@Dr15Jones, @cmsbuild, @civanch, @subirsarkar, @mdhildreth, @bsunanda, @makortel, @srimanob can you please review it and eventually sign? Thanks.
@bsunanda this is something you requested to watch as well.
@sextonkennedy, @antoniovilela, @rappoccio you are the release manager for this.

cms-bot commands are listed here

[fabiocos]

type bugfix

[fabiocos]

test parameters:

• workflows = 29634.0,29634.911
• relvals_opt = --what upgrade

[fabiocos]

please test

[cmsbuild]

-1

Failed Tests: RelVals
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-f8963a/39858/summary.html
COMMIT: 9468e88
CMSSW: CMSSW_14_1_X_2024-06-12-1900/el8_amd64_gcc12
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/45211/39858/install.sh to create a dev area with all the needed externals and cmssw changes.

RelVals

• 29634.911A fatal system signal has occurred: segmentation violation

[fabiocos]

@srimanob as far as I can see, the residual crash is related to HGCAL, the MTD problem is solved. I will update the standalone geometry tests to bring them to D110 as default, but the problem anyway was not present there, it is the way Geant4 is passing the information that causes the issue. I need to check whether I can update the test to catch that.

[jfernan2]

+1

[srimanob]

The workflow 29634.911 is success now, thanks @fabiocos

https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-17a93c/39885/runTheMatrix-results/29634.911_TTbar_14TeV+2026D110_DD4hep/

[srimanob]

+Upgrade

[civanch]

+1

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @rappoccio, @antoniovilela, @sextonkennedy (and backports should be raised in the release meeting by the corresponding L2)
Notice This PR was tested with additional Pull Request(s), please also merge them if necessary: cms-data/Geometry-TestReference#14

[fabiocos]

@cms-sw/orp-l2 this PR requires to be merged together with cms-data/Geometry-TestReference#14 , otherwise several unit tests will fail.

[rappoccio]

+1

[aandvalenzuela]

Hi @fabiocos,

Unit test RecoMTDDetLayersTestDriver fails in aarch64 with:

>  L/W/T   :     118.869904 /     237.800034 /       0.005005
201367c201367
<  rmax    :     118.900063
---
>  rmax    :     118.900070
Failure in comparison for mtdDetLayerGeometry.log: status 255

---> test RecoMTDDetLayersTestDriver had ERRORS
TestTime:69
^^^^ End Test RecoMTDDetLayersTestDriver ^^^^

I think the failure might be related with this PR. See full log in CMSSW_14_1_X_2024-06-17-2300.

Thanks,
Andrea

[fabiocos]

@aandvalenzuela thanks for warning me. We are running this test since years now, and this is the first issue of this type as far as I recall. It looks like a problem with numerical rounding on aarch64 vs amd64 that was not so far exposed. Apart that I have no easy direct way to test on aarch64, I do not see really what I could do, apart for reducing the number of significant digits to be checked, reproduce the reference and hoping for the best. Essentially, the message from the failure above is that we are unable to rely on geometrical calculations at better than 1 micron (the values causing issues are positions expressed in cm).

[fwyzard]

@fabiocos you can use lxplus-arm.cern.ch to run tests on aarch64.

[fabiocos]

@fwyzard thanks for the suggestion, it has been useful