Add workflows for profiling the GPU code

[fwyzard]

PR description:

Add four workflows for profiling the GPU code:

• .504 Pixel-only local reconstruction and quadruplets
• .508 Pixel-only local reconstruction and triplets
• .514 ECAL-only local reconstruction
• .524 ECAL-only local reconstruction

The workflows explicitly consume the GPU products, so they can only run on a GPU-equipped machine.

The transfer to the host and the conversion to the legacy format is not run.

PR validation:

Used to profile the various workflows on top of CMSSW_12_1_0_pre3, runing over that release's TTbar relvals with pileup:

measurement	CMSSW_12_1_0_pre3
I/O throughput	~ 2 kev/s
11634.504	1071 ± 3 ev/s
11634.508	560 ± 2 ev/s
11634.514	1391 ± 5 ev/s
11634.524	1354 ± 10 ev/s

[fwyzard]

please test

[fwyzard]

no need to test explicitly on GPU, since the new workflows are not part of any test

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-35540/25762

• This PR adds an extra 24KB to repository

• There are other open Pull requests which might conflict with changes you have proposed:

  • File Configuration/PyReleaseValidation/python/upgradeWorkflowComponents.py modified in PR(s): Add run3 tracking low pu era #33532, Phase 2 HLT menu targetting 7.5×10³⁴ cm⁻²s⁻¹ #35342, Add NanoAOD DQM to RelVal wfs #35412

[cmsbuild]

A new Pull Request was created by @fwyzard (Andrea Bocci) for master.

It involves the following packages:

• Configuration/PyReleaseValidation (pdmv, upgrade)

@jordan-martins, @bbilin, @wajidalikhan, @AdrianoDee, @srimanob, @kskovpen can you please review it and eventually sign? Thanks.
@makortel, @kpedro88, @Martin-Grunewald, @missirol, @fabiocos, @slomeo this is something you requested to watch as well.
@perrotta, @dpiparo, @qliphy you are the release manager for this.

cms-bot commands are listed here

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-04c13f/19411/summary.html
COMMIT: b5262fc
CMSSW: CMSSW_12_1_X_2021-10-05-1100/slc7_amd64_gcc900
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/35540/19411/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:

• @kbunkow new package L1Trigger/L1TMuonOverlapPhase1 (new version of the L1 Trigger Overlap Muon Track Finder (OMTF) emulator) #35111

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-04c13f/19411/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-04c13f/19411/git-merge-result

• DAS Queries: The DAS query tests failed, see the summary page for details.

Comparison Summary

Summary:

• No significant changes to the logs found
• Reco comparison results: 16 differences found in the comparisons
• DQMHistoTests: Total files compared: 40
• DQMHistoTests: Total histograms compared: 3219394
• DQMHistoTests: Total failures: 94
• DQMHistoTests: Total nulls: 1
• DQMHistoTests: Total successes: 3219277
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: -0.004 KiB( 39 files compared)
• DQMHistoSizes: changed ( 312.0 ): -0.004 KiB MessageLogger/Warnings
• Checked 169 log files, 37 edm output root files, 40 DQM output files
• TriggerResults: no differences found

[fwyzard]

ping @cms-sw/pdmv-l2

[fwyzard]

ping @cms-sw/upgrade-l2

[srimanob]

+Upgrade

This PR is to add 4 workflows for profiling.

[kskovpen]

+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. @perrotta, @dpiparo, @qliphy (and backports should be raised in the release meeting by the corresponding L2)

[perrotta]

+1

[iarspider]

Hi @fwyzard ,

we observe RelVal failures in GPU_X builds starting Friday: https://cmssdt.cern.ch/SDT/html/cmssdt-ib/#/relVal/CMSSW_12_1/2021-10-10-2300?selectedArchs=slc7_amd64_gcc900&selectedFlavors=GPU_X&selectedStatus=failed

Could these be related to this PR?

Thanks!

[fwyzard]

Most likely it is due to #35566 .
Should be fixed by #35630 .