Update NNVtx emulator to match latest HW

[rtschmitz]

PR description:

This PR adds the following changes to NNvtx models:

• Weight, pattern, and association tensorflow models have been replaced by hls4ml, including updates to emulator to match model quantization of QKeras training
• NN weight replaces pT in the output VertexWord when running NN vertexing (before: only vertex field filled)
• In case the output of the NNvtx location results in a tie between two bins, a new tiebreaking procedure has been developed. Before, an average was taken between each of the tied bins. Now, the most central bin is chosen in the case of a tie.

PR validation:

HW tests have shown 100% agreement with this new emulation. The new vertex bin tiebreak procedure has shown equal or better performance in emulation, with better timing performance on HW. Results have been presented in a recent GTT meeting.

[cmsbuild]

cms-bot internal usage

[cmsbuild]

-code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49430

ERROR: Build errors found during clang-tidy run.

src/L1Trigger/VertexFinder/src/VertexFinder.cc:1468:35: error: use of overloaded operator '>=' is ambiguous (with operand types 'ap_fixed<10, 1>' and 'const float') [clang-diagnostic-error]
 1468 |       nnOutput[i] = (vx_output[i] >= histogrammingThreshold_) ? vx_output[i].to_float() : 0.0f;
      |                      ~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~~~~~~~~~
/cvmfs/cms-ib.cern.ch/sw/x86_64/nweek-02942/el8_amd64_gcc13/external/hls/2025.05-1e9bc123eaf0d9d12de4203b234629e6/include/ap_fixed_base.h:1478:3: note: candidate function
--
gmake: *** [config/SCRAM/GMake/Makefile.coderules:129: code-checks] Error 2
gmake: *** [There are compilation/build errors. Please see the detail log above.] Error 2

[cmsbuild]

-code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49453

• There are other open Pull requests which might conflict with changes you have proposed:
  • File L1Trigger/DemonstratorTools/test/gtt/createFirmwareInputFiles_cfg.py modified in PR(s): W3Pi Emulation and Test Vectors #49003, Gtt jet finding geometry readout #50908

Code check has found code style and quality issues which could be resolved by applying following patch(s)

• code-checks:
  https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49453/code-checks.patch
  e.g. curl -k https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49453/code-checks.patch | patch -p1
  You can also run scram build code-checks to apply code checks directly

• code-format:
  https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49453/code-format.patch
  e.g. curl -k https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49453/code-format.patch | patch -p1
  You can also run scram build code-format to apply code format directly

[cmsbuild]

-code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49454

Code check has found code style and quality issues which could be resolved by applying following patch(s)

• code-checks:
  https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49454/code-checks.patch
  e.g. curl -k https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49454/code-checks.patch | patch -p1
  You can also run scram build code-checks to apply code checks directly

• code-format:
  https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49454/code-format.patch
  e.g. curl -k https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49454/code-format.patch | patch -p1
  You can also run scram build code-format to apply code format directly

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51008/49456

[cmsbuild]

Pull request #51008 was updated.

[quinnanm]

please test

[cmsbuild]

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

It involves the following packages:

• L1Trigger/L1TTrackMatch (l1)
• L1Trigger/VertexFinder (l1)

@BenjaminRS, @quinnanm can you please review it and eventually sign? Thanks.
@Martin-Grunewald, @missirol, @mmusich this is something you requested to watch as well.
@ftenchini, @mandrenguyen, @sextonkennedy you are the release manager for this.

cms-bot commands are listed here

[cmsbuild]

-1

Failed Tests: RelVals
Size: This PR adds an extra 16KB to repository
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-821d1e/53510/summary.html
COMMIT: 495734a
CMSSW: CMSSW_17_0_X_2026-05-27-1100/el8_amd64_gcc13
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/51008/53510/install.sh to create a dev area with all the needed externals and cmssw changes.

Failed RelVals

• 34434.911

----- Begin Fatal Exception 27-May-2026 19:54:52 CEST-----------------------
An exception of category 'StdException' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=VertexProducer label='l1tVertexFinder'
Exception Message:
A std::exception was thrown.
hls4ml emulator model library dlopen failure!
----- End Fatal Exception -------------------------------------------------

• 34434.0

----- Begin Fatal Exception 27-May-2026 19:50:14 CEST-----------------------
An exception of category 'StdException' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=VertexProducer label='l1tVertexFinder'
Exception Message:
A std::exception was thrown.
hls4ml emulator model library dlopen failure!
----- End Fatal Exception -------------------------------------------------

• 34434.75

----- Begin Fatal Exception 27-May-2026 19:50:31 CEST-----------------------
An exception of category 'StdException' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=VertexProducer label='l1tVertexFinder'
Exception Message:
A std::exception was thrown.
hls4ml emulator model library dlopen failure!
----- End Fatal Exception -------------------------------------------------

Expand to see more relval errors ...

• 34634.999
• 34500.0
• 34496.0

[cmsbuild]

Milestone for this pull request has been moved to CMSSW_20_0_X. Please open a backport if it should also go in to CMSSW_17_0_X.

[makortel]

Just checking, is this PR intended for Phase 2 (20_1_X) or Run 3 legacy (17_0_X) or both?

[rtschmitz]

This is for Phase 2; it's an update to a vertexing algorithm related to the L1 Tracker upgrade. So I think it should be for 20_1_X, but I'll let @BenjaminRS confirm.

[makortel]

Thanks. For the case of 20_1_X-only nothing needs to be done.

[BenjaminRS]

Hi Matti and Ryan - indeed we will just have this Phase 2 PR going to 20_1_X.

[cmsbuild]

Milestone for this pull request has been moved to CMSSW_20_1_X. Please open a backport if it should also go in to CMSSW_20_0_X.