First Round of Phase 1 Tracking changes for Heavy Ions

[abaty]

This PR makes many changes aimed at improving the Heavy Ion Tracking for use with the Phase 1 pixel detector. In particular, it is for HI tracking done under the Run2_2017_trackingPhase1QuadProp era. The following changes are made:

1. Fixed a missing product in the regional muon tracking which was causing a crash.
2. Changed tracking iterations to use all 4 layers of the pixels instead of the first 3
3. Added tracking iterations: LowPtQuadStep, DetachedQuadStep, and highPtTripletStep
4. Changed seeding from triplets to Cellular Automaton quadruplets

Most of these changes are attached to the 'trackingPhase1QuadProp.' modifier. I am not sure if this is more appropriate to use than the 'trackingPhase1' modifer; perhaps someone can comment.

In order to Test this PR, one needs do the following:

1. Merge the PR and do a 'scram build.'
2. Comment out lines 38-50 in RecoHI/Configuration/python/Reconstruction_HI_cff.py in order to prevent issues related to workflows downstream of the tracking (some of these do not work in Run2_2017 era, but are out of the scope of this PR).
3. Copy and run the config file found here: /afs/cern.ch/user/a/abaty/work/public/forPeople/forPhase1PR_May5/
   , which runs the Run2_2017 track reconstruction on a sample of MB Hydjet events privately produced with SIM of the new pixel detector.
   @mandrenguyen

[cmsbuild]

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

It involves the following packages:

RecoHI/Configuration
RecoHI/HiMuonAlgos
RecoHI/HiTracking

@perrotta, @cmsbuild, @slava77, @davidlange6 can you please review it and eventually sign? Thanks.
@MiheeJo, @jazzitup, @echapon, @yenjie, @kurtejung, @mandrenguyen, @dgulhan, @yetkinyilmaz this is something you requested to watch as well.
@davidlange6 you are the release manager for this.

cms-bot commands are listed here

[slava77]

@cmsbuild please test

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/19700/console Started: 2017/05/09 21:06

[cmsbuild]

+1
Tested at: 9338fc0
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-18646/19700/summary.html

[cmsbuild]

Comparison job queued.

[cmsbuild]

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-18646/19700/summary.html

Comparison Summary:

• You potentially added 1559 lines to the logs
• Reco comparison results: 1848 differences found in the comparisons
• DQMHistoTests: Total files compared: 24
• DQMHistoTests: Total histograms compared: 1834100
• DQMHistoTests: Total failures: 26882
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 1807038
• DQMHistoTests: Total skipped: 180
• DQMHistoTests: Total Missing objects: 0
• Checked 98 log files, 14 edm output root files, 24 DQM output files

[cmsbuild]

Pull request #18646 was updated. @perrotta, @cmsbuild, @slava77, @davidlange6 can you please check and sign again.

[abaty]

After discussions with @mandrenguyen , we have decided to add an additional commit which turns off the MVA track selectors in the HI tracking workflow. We will turn them back on once new ones can be trained later in the development cycle.

[perrotta]

please test

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/19741/console Started: 2017/05/10 23:19

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/20166/console Started: 2017/05/29 16:30

[perrotta]

@abaty : do you also plan to provide a "2018 phase1" HI MC workflow to test more easily these and forthcoming HI developments? That would be really welcome: it could either stay in this PR, or in a separate one if you prefer.

[abaty]

@perrotta : I will defer to @mandrenguyen 's decision with regards to a 2018 workflow. Regardless, I would prefer place it in a separate PR if we decide to provide one now.

[mandrenguyen]

I'll add a phase I MC wf in a separate PR once this one is merged.

…

Sent from my iPhone

On May 29, 2017, at 16:49, abaty ***@***.***> wrote: @perrotta : I will defer to @mandrenguyen 's decision with regards to a 2018 workflow. Regardless, I would prefer place it in a separate PR if we decide to provide one now. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[cmsbuild]

+1
Tested at: 193f14e
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-18646/20166/summary.html

[cmsbuild]

Comparison job queued.

[cmsbuild]

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-18646/20166/summary.html

Comparison Summary:

• You potentially added 197 lines to the logs
• Reco comparison results: 3413 differences found in the comparisons
• DQMHistoTests: Total files compared: 24
• DQMHistoTests: Total histograms compared: 1840355
• DQMHistoTests: Total failures: 48614
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 1791561
• DQMHistoTests: Total skipped: 180
• DQMHistoTests: Total Missing objects: 0
• Checked 98 log files, 14 edm output root files, 24 DQM output files

[makortel]

@abaty In addition, depending on your further plans, it might be more clear to change the hiRegitMuInitialStepSeedLayers to have triplet layers instead of quadruplets. On the other hand, the doublet and triplet generators work exactly the same whether the input seeding layers are phase0 initialStep triplets or phase1 initialStep quadruplets, so this would be more of a cosmetic change (one source of confusion less when debugging by looking edmConfigDump output).

[perrotta]

After the last commit, there are no more crashes in my tests jobs,

I also looked at the cpu time taken by the new tracking modules: although it is not easy to compare with workflows running with a different track reconstruction (as it is the case for both phase1 pp or pre-phase1 HI reco), I don't see anything worrying. The same for the size of the produced track objects, which does not show any evident blow up. Of course, some better comparison will be likely performed as soon as some dedicated workflow will be prepared, as anticipated in #18646 (comment)

At this point, this PR is basically ready to be signed off by reco. Just let us know if you plan to follow up with the suggestion of @makortel in #18646 (comment) , or if you would like to also postpone it to a future PR.

[abaty]

@makortel @perrotta We are planning on making a larger PR fairly soon that revamps to muon tracking to have full or nearly-full parity with the pp algorithms. The people who bid EPR to do this sit across from me in my office, and are aware of the changes made here. If we moved from quad to triplet seed layers, I would expect that change to be undone very soon, so I don't think it is necessary at this point.

I will just also mention that we have done preliminary timing tests as a function of centrality, and did not see anything too shocking when comparing with the old tracking times.

[makortel]

If we moved from quad to triplet seed layers, I would expect that change to be undone very soon, so I don't think it is necessary at this point.

I agree.

[perrotta]

+1
According to #18646 (comment) and clarification in #18646 (comment)

[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 requires discussion in the ORP meeting before it's merged. @davidlange6, @smuzaffar

[davidlange6]

+1