Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Bug Fix for CLUE in BH section of HGCAL #35727

Merged
merged 1 commit into from Oct 20, 2021
Merged

Conversation

rovere
Copy link
Contributor

@rovere rovere commented Oct 19, 2021

PR description:

Possibly for historical reasons dating back to the original
implementation of the imaging algorithm, from which CLUE has been
derived, the critical distance in the BH section is treated differently.
There is no physics driven motivation for this behaviour. This sets
the critical distance parameter equal all over the Silicon detectors
of HGCAL. The effect is visible in a much better separation of sibling
showers in the BH section.

PR validation:

Validated on a privately produced sample of two CloseBy photons in BH separated by 3.5cm to test merging vs separation.

Before the fix

BH43_Evt15_wrong_deltac

After the fix

BH43_Evt15_correct_deltac

Possibly for historical reasons dating back to the original
implementation of the _imaging_ algorithm, from which CLUE has been
derived, the critical distance in the BH section is treated differently.
There is not _physics driven motivation_ for this behaviour. This set
the critical distance parameter identical all over the Silicon detectors
of HGCAL. The effect is visible in a much better separation of sibling
showers in the BH section.
@cmsbuild
Copy link
Contributor

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-35727/26052

  • This PR adds an extra 16KB to repository

@cmsbuild
Copy link
Contributor

A new Pull Request was created by @rovere (Marco Rovere) for master.

It involves the following packages:

  • RecoLocalCalo/HGCalRecProducers (upgrade, reconstruction)

@jpata, @AdrianoDee, @srimanob, @cmsbuild, @slava77 can you please review it and eventually sign? Thanks.
@edjtscott, @vandreev11, @sethzenz, @bsunanda, @felicepantaleo, @lgray, @cseez, @apsallid, @pfs, @lecriste, @hatakeyamak, @ebrondol, @clelange 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

@rovere
Copy link
Contributor Author

rovere commented Oct 19, 2021

assign hgcal-dpg

@rovere
Copy link
Contributor Author

rovere commented Oct 19, 2021

@cmsbuild please test

@cmsbuild
Copy link
Contributor

New categories assigned: hgcal-dpg

@felicepantaleo,@rovere,@pfs,@cseez you have been requested to review this Pull request/Issue and eventually sign? Thanks

@rovere
Copy link
Contributor Author

rovere commented Oct 19, 2021

type bug-fix

@slava77
Copy link
Contributor

slava77 commented Oct 19, 2021

@rovere
do you have some plots for the impact on physics quantities (e.g. response or resolution)?
I'm not sure if the short matrix tests are enough to record the impact if it's significant.

@rovere
Copy link
Contributor Author

rovere commented Oct 19, 2021

@slava77
thanks for asking. I think it will be rather difficult to appreciate the impact of the fix by looking at the resolution.
The resolution in that region of the detector using reconstructed objects is already in the 20% ballpark, far from being optimal. Which means, as a consequence, that also the Pattern Recognition could be improved.
The response should not be affected by this fix: we were not "loosing energy", we were just merging clusters together.

What I could offer is a comparison of the fake rate for the layer clusters (vs layer number) computed on the special sample I mentioned above:

  • red is this PR
  • blue is baseline

Screenshot_20211019_151915

@slava77
Copy link
Contributor

slava77 commented Oct 20, 2021

@cmsbuild please test

it seems like the last attempt got stuck.
I had another PR test done already under 5 hours, while this one is 17hrs in.

@cmsbuild
Copy link
Contributor

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-c094c1/19755/summary.html
COMMIT: 2dfe916
CMSSW: CMSSW_12_1_X_2021-10-19-1100/slc7_amd64_gcc900
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/35727/19755/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

  • No significant changes to the logs found
  • Reco comparison results: 3417 differences found in the comparisons
  • DQMHistoTests: Total files compared: 40
  • DQMHistoTests: Total histograms compared: 2751113
  • DQMHistoTests: Total failures: 14913
  • DQMHistoTests: Total nulls: 0
  • DQMHistoTests: Total successes: 2736178
  • DQMHistoTests: Total skipped: 22
  • DQMHistoTests: Total Missing objects: 0
  • DQMHistoSizes: Histogram memory added: 0.0 KiB( 39 files compared)
  • Checked 170 log files, 37 edm output root files, 40 DQM output files
  • TriggerResults: no differences found

@rovere
Copy link
Contributor Author

rovere commented Oct 20, 2021

Of the many (small) differences introduced by this fix, I'd pick the following as the most important, IMHO:
FixDeltaC

This shows that the average (profile) number of layer clusters associated with a TICL-SimTrackster is now flat and does not show any longer the sudden drop at the layer at which we had the wrong delta_c parameter configured.

I'll let others, if interested, review and comment further.
As far as I could see from private tests (linked at the top of this PR), and from inspection of the PR differences, I'm happy with this fix.

@rovere
Copy link
Contributor Author

rovere commented Oct 20, 2021

+1

@slava77
Copy link
Contributor

slava77 commented Oct 20, 2021

+reconstruction

for #35727 2dfe916

  • code changes are in line with the PR description
  • jenkins tests pass and comparisons with the baseline show differences in phase-2 workflows; looking at e.g. 34834.999 the differences start in hgcalLayerClusters and propagate downstream. More validation plots were linked earlier in the thread.

@AdrianoDee
Copy link
Contributor

+upgrade

@cmsbuild
Copy link
Contributor

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
Copy link
Contributor

+1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

5 participants