Update mkFit-related processModifier for LST-seeded phase-2 HLT single tracking iteration

[mmasciov]

PR description:

As per title, for upcoming update of the HLT phase-2 tracking baseline, building on:

• CA extension to OT + LST for track seeding;
• mkFit for track building.

For more details, this follows previous work and discussions; see presentation by @bdanzi.

PR validation:

See presentation.
Re-validated in 160X, obtaining analogous results (see recap below).

Can be tested with workflow 29634.7571, introduced in #49231.

NOTE: it requires cms-data/RecoTracker-MkFit#19

Recap of physics performance:

Blue: current baseline, with two iterations using patatrack IT-only ntuples (>=4 hits) + legacy triplets as seeds, respectively;
Red: previous baseline, with two iterations using legacy quadruplets + legacy triplets as seeds, respectively - procModifiers=[phase2LegacyPixelTracks];
Black: future baseline, with a single iteration using seeds resulting from CA OT-extension + LST, and mkFit for track building - procModifiers=[phase2CAExtension,singleIterPatatrack,trackingLST,seedingLST,trackingMkFitCommon,hltTrackingMkFitInitialStep];
Orange: for comparison, single iteration using seeds resulting from CA OT-extension + LST, and CKF for track building - procModifiers=[phase2CAExtension,singleIterPatatrack,trackingLST,seedingLST].

Focusing on physics performance for the best single-iteration approach (in black, using mkFit):

• Efficiency is comparable to current baseline, with mild inefficiency at 2 < |η| < 3 (see presentation for more details);
• Fake+duplicate rate is significantly reduced.
• Using mkFit for track building instead of CKF allows to improve both efficiency and fake+duplicate rate, as well as timing.

For timing of (pixel) tracking, i.e., summing all (pixel) tracking modules from usual circles, from single-thread measurements - using GPUs:

• Current baseline = X (X~2 s)
• Previous baseline = ~1.3X
• Single iteration (CA extension+LST seeds) with mkFit (future baseline) = ~0.25X
• Single iteration (CA extension+LST seeds) with CKF = ~0.35X
  (--> speed-up of full tracking by ~1.4X when mkFit is used for track building wrt. CKF).

For timing of (pixel) tracking, i.e., summing all (pixel) tracking modules from usual circles, from single-thread measurements - using CPU-only, for completeness:

• Current baseline = X (X~2.7 s)
• Previous baseline = ~1.075X
• Single iteration (CA extension+LST seeds) with mkFit (future baseline) = ~0.99X (fastest option also for CPU-only)
• Single iteration (CA extension+LST seeds) with CKF = ~1.09X

[cmsbuild]

cms-bot internal usage

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-49383/46808

[cmsbuild]

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

It involves the following packages:

• HLTrigger/Configuration (hlt)

@Martin-Grunewald, @cmsbuild, @mmusich can you please review it and eventually sign? Thanks.
@Martin-Grunewald, @SohamBhattacharya, @VourMa, @missirol, @mmusich, @rovere 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

[mmusich]

test parameters:

• pull_request = Add JSON for LST-seeded iteration (phase-2 HLT) cms-data/RecoTracker-MkFit#19
• enable = hlt_p2_integration, hlt_p2_timing
• workflows = ph2_hlt

[mmusich]

@cmsbuild, please test

[mmusich]

Blue: current baseline, with two iterations using patatrack IT-only ntuples (>=4 hits) + legacy triplets as seeds, respectively;
Red: previous baseline, with two iterations using legacy quadruplets + legacy triplets as seeds, respectively;
Black: future baseline, with a single iteration using seeds resulting from CA OT-extension + LST, and mkFit for track building;
Orange: for comparison, single iteration using seeds resulting from CA OT-extension + LST, and CKF for track building.

for my own book-keeping / understanding, may I ask you to which combination of process modifiers as currently available in release each of these curves corresponds to?

[cmsbuild]

+1

Size: This PR adds an extra 44KB to repository
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-6ccbc5/49435/summary.html
COMMIT: c90e642
CMSSW: CMSSW_16_0_X_2025-11-12-2300/el8_amd64_gcc13
Additional Tests: HLT_P2_INTEGRATION,HLT_P2_TIMING
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/49383/49435/install.sh to create a dev area with all the needed externals and cmssw changes.

The following merge commits were also included on top of IB + this PR after doing git cms-merge-topic:

• @dan131riley disable ubsan checks for some global summary functions #49301
• @fwyzard Fix BuildFiles, includes and comments #49374
• @Dr15Jones Fix UBSAN error in FWCore/Framework unit tests #49370
• @iarspider TrackerGeometry: add assert() to offsetDU and endsetDU #49366
• @soureek Updating fillDescriptions in RecoBTag area #49379

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-6ccbc5/49435/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-6ccbc5/49435/git-merge-result

HLT P2 Timing: chart

Comparison Summary

Summary:

• You potentially added 14 lines to the logs
• ROOTFileChecks: Some differences in event products or their sizes found
• Reco comparison results: 6 differences found in the comparisons
• Reco comparison had 2 failed jobs
• DQMHistoTests: Total files compared: 68
• DQMHistoTests: Total histograms compared: 4352785
• DQMHistoTests: Total failures: 564
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 4352201
• DQMHistoTests: Total skipped: 20
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 67 files compared)
• Checked 275 log files, 231 edm output root files, 68 DQM output files
• TriggerResults: no differences found

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-49383/46827

• There are other open Pull requests which might conflict with changes you have proposed:
  • File Configuration/PyReleaseValidation/README.md modified in PR(s): [mkFit] final fit initial PR and testing workflow #49339
  • File Configuration/PyReleaseValidation/python/relval_Run4.py modified in PR(s): [mkFit] final fit initial PR and testing workflow #49339
  • File Configuration/PyReleaseValidation/python/upgradeWorkflowComponents.py modified in PR(s): [mkFit] final fit initial PR and testing workflow #49339
  • File Configuration/PyReleaseValidation/scripts/runTheMatrix.py modified in PR(s): [Do not merge] For cms-bot PR testing #48637, [mkFit] final fit initial PR and testing workflow #49339

[cmsbuild]

Pull request #49383 was updated. @AdrianoDee, @DickyChant, @Martin-Grunewald, @antoniovagnerini, @cmsbuild, @miquork, @mmusich can you please check and sign again.

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-49383/46867

• There are other open Pull requests which might conflict with changes you have proposed:
  • File Configuration/PyReleaseValidation/python/upgradeWorkflowComponents.py modified in PR(s): Fix location of alpaka modification for particleFlowClusterHCAL #49295, ECAL Phase 2 - Upgrades for Phase 2 development workflow, validation and DQM #49409
  • File Configuration/PyReleaseValidation/scripts/runTheMatrix.py modified in PR(s): [Do not merge] For cms-bot PR testing #48637

[cmsbuild]

Pull request #49383 was updated. @AdrianoDee, @DickyChant, @antoniovagnerini, @cmsbuild, @miquork can you please check and sign again.

[bdanzi]

@mmasciov unfortunately now this has conflicts that must be resolved.

It should be ok now, I took the occasion to introduce the Phase2CAExtension modifier also in .7572 variant workflow, please ask for testing it

[slava77]

@cmsbuild please test

[cmsbuild]

+1

Size: This PR adds an extra 72KB to repository
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-6ccbc5/49533/summary.html
COMMIT: 226447f
CMSSW: CMSSW_16_0_X_2025-11-18-1100/el8_amd64_gcc13
Additional Tests: HLT_P2_INTEGRATION,HLT_P2_TIMING
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmssw/49383/49533/install.sh to create a dev area with all the needed externals and cmssw changes.

The following merge commits were also included on top of IB + this PR after doing git cms-merge-topic:

• @denizsun [HCAL-DQM] Additional plots: mask channels in HEP07 (RM 3, RMfib 4) affected by sporadic bad data #49394
• @cms-sw Update tag for RecoLocalCalo-HGCalRecProducers to V00-03-00 cmsdist#10209

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-6ccbc5/49533/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-6ccbc5/49533/git-merge-result

HLT P2 Timing: chart

Comparison Summary

Summary:

• You potentially added 12 lines to the logs
• ROOTFileChecks: Some differences in event products or their sizes found
• Reco comparison results: 8 differences found in the comparisons
• DQMHistoTests: Total files compared: 69
• DQMHistoTests: Total histograms compared: 4399383
• DQMHistoTests: Total failures: 5087
• DQMHistoTests: Total nulls: 8
• DQMHistoTests: Total successes: 4394268
• DQMHistoTests: Total skipped: 20
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 68 files compared)
• Checked 278 log files, 233 edm output root files, 69 DQM output files
• TriggerResults: found differences in 2 / 67 workflows

[AdrianoDee]

+pdmv

[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. @sextonkennedy, @mandrenguyen, @ftenchini (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/RecoTracker-MkFit#19

[ftenchini]

+1