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

Change default radii for HGCAL layer and multi clustering #20131

Merged
merged 1 commit into from Aug 18, 2017

Conversation

edjtscott
Copy link
Contributor

Would like to change the default values for two sets of radii in the HGCAL clustering:

  1. the delta_c values in the layer clustering (from 2,2,2 to 2,2,5)
  2. the multiclustering radii (from 2,2,2 to 2,5,5)

This makes more sense for hadronic objects (eg in the BH the cell sizes can be more than 2cm so the 2cm delta_c is silly) but is expected to have almost no effect on EM objects.

See plots here and here for the default and updated values respectively for unconverted pT 25 photons.

@cmsbuild cmsbuild changed the base branch from CMSSW_9_3_X to master August 11, 2017 14:04
@cmsbuild
Copy link
Contributor

@edjtscott, CMSSW_9_3_X branch is closed for direct updates. cms-bot is going to move this PR to master branch.
In future, please use cmssw master branch to submit your changes.

@cmsbuild
Copy link
Contributor

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

It involves the following packages:

RecoLocalCalo/HGCalRecProducers

@perrotta, @cmsbuild, @kpedro88, @slava77 can you please review it and eventually sign? Thanks.
@vandreev11, @sethzenz, @felicepantaleo, @kpedro88, @lgray, @cseez, @pfs this is something you requested to watch as well.
@davidlange6 you are the release manager for this.

cms-bot commands are listed here

@cmsbuild
Copy link
Contributor

@slava77
Copy link
Contributor

slava77 commented Aug 11, 2017

@cmsbuild please test

@felicepantaleo @rovere
please comment on relevance of this for pre4/TDR production.

@cmsbuild
Copy link
Contributor

cmsbuild commented Aug 11, 2017

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/22222/console Started: 2017/08/11 16:15

@slava77
Copy link
Contributor

slava77 commented Aug 11, 2017

assign upgrade

@rovere
Copy link
Contributor

rovere commented Aug 11, 2017

@slava77 this will only affect objects derived from multi-clustering within HGCal, namely the ones activated by #20045
In fact, the changes are mostly for the BH part of HGCal, while we only build em objects that will be confined in EE. So, to summarize, it's nice to have, will have negligible effects on the objects produced in #20045 but would be better for hadronic objects that, at some point, will come too.

@cmsbuild
Copy link
Contributor

@cmsbuild
Copy link
Contributor

Comparison job queued.

@cmsbuild
Copy link
Contributor

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-20131/22222/summary.html

Comparison Summary:

  • No significant changes to the logs found
  • Reco comparison results: 22 differences found in the comparisons
  • DQMHistoTests: Total files compared: 26
  • DQMHistoTests: Total histograms compared: 2713569
  • DQMHistoTests: Total failures: 246
  • DQMHistoTests: Total nulls: 0
  • DQMHistoTests: Total successes: 2713134
  • DQMHistoTests: Total skipped: 189
  • DQMHistoTests: Total Missing objects: 0
  • Checked 107 log files, 14 edm output root files, 26 DQM output files

@slava77
Copy link
Contributor

slava77 commented Aug 17, 2017

+1

for #20131 cd9607d

  • increased clustering size in BH and multiclustering size in FH and BH parts of HGCAL so that the hadronic part is accumulated in larger volume
  • jenkins tests pass and reco comparisons with the (5 days old) baseline show changes only in 2023 workflows in hgcalLayerClusters
  • local tests with PU200 on top of Add MultiCluster derived electrons and photons #20045 show somewhat expected behavior: changes in the hgcalLayerClusters propagate downstream to the newly added FromMultiCl collections
    • number of multiclusters is down by 7% mostly due to migration and accumulation in hadron parts of HGCAL
      all_sign955-2023pu200vssign951-pass-10c3208-2023pu200_ttbar14tev2023d17putimingwf20234p2c_log10recohgcalmulticlusters_hgcallayerclusters__reco_obj_energy
    • this is seen as a 15% drop in the number of PF clusters "FromMultiCl"
      all_sign955-2023pu200vssign951-pass-10c3208-2023pu200_ttbar14tev2023d17putimingwf20234p2c_log10recopfclusters_particleflowclusterhgcalfrommulticl__reco_obj_energy
    • and given higher thresholds, there is a 27% increase in the number of PF superclusters
      all_sign955-2023pu200vssign951-pass-10c3208-2023pu200_ttbar14tev2023d17putimingwf20234p2c_log10recosuperclusters_particleflowsuperclusterhgcalfrommulticl__reco_obj_energy
    • more energy in hadron part is apparently picked up in electron seeding as larger h/e and we get less gsfTracks and less gsfElectrons (still, less by only 7%)
      all_sign955-2023pu200vssign951-pass-10c3208-2023pu200_ttbar14tev2023d17putimingwf20234p2c_recogsfelectrons_ecaldrivengsfelectronsfrommulticl__reco_obj_pt
      .. oddly enough this turns around as a slight increase in the number of photons
      all_sign955-2023pu200vssign951-pass-10c3208-2023pu200_ttbar14tev2023d17putimingwf20234p2c_recophotons_photonsfrommulticl__reco_obj_caloposition_eta

On the technical side:

  • the event size is about the same in FEVT or RECO: more PF superclusters are balanced out by fewer GSF tracks and electrons
  • CPU looks about the same: there is perhaps 1% decrease of the total reco from changes in "FromMultiCl" modules which have a mostly reduced CPU use.

@davidlange6
Copy link
Contributor

+1

@slava77
Copy link
Contributor

slava77 commented Aug 18, 2017

@davidlange6
you will need to "merge" to get this in. "+1" is not enough.

@davidlange6
Copy link
Contributor

merge

@cmsbuild cmsbuild merged commit d0b34bf into cms-sw:master Aug 18, 2017
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