ele/ph mass fix #36889

rgoldouz · 2022-02-04T17:18:05Z

PR description:

Hello,
The electron and photon mass are not saved properly and nanoAOD users reported that many times. It is not related to nanoAOD. The problem is available in miniAOD too.
You can check the printout here
http://dalfonso.web.cern.ch/dalfonso/NANO/massDump.txt

I checked and found that the problem is not from the EGamma code but from TLorentzVector root class. I found that the TLorentz vector is set via pxpypzE for all object candidates. It seems fine but when I tested locally for a particle with pxpypzE=(1,1,1,sqrt(3)) I did not get the mass==0. You can easily reproduce it
root -l
root [1] typedef ROOT::Math::LorentzVector<ROOT::Math::PxPyPzE4D > XYZTLorentzVectorD
root [2] XYZTLorentzVectorD newP4(1,1,1,sqrt(3));
root [3] newP4.M()
(double) -2.1073424e-08

It is because of the precision problem in TLorentz vector and it is recommended to use PxPyPzM to have correct masses for low mass objects.
https://root.cern.ch/doc/master/classROOT_1_1Math_1_1PxPyPzM4D.html

So I set the electron and photon masses by hand to 0.0005 and 0 GeV , respectively.

PR validation:

I generated electron and photon up to reco level using fastsim cmsDriver command and checked the photon and electron mass before and after this fix. Wrong mass values before the fix are corrected after the fix.

please let me know if you have any question.
Thanks ,
Reza

cmsbuild · 2022-02-04T17:26:40Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-36889/28160

This PR adds an extra 32KB to repository

cmsbuild · 2022-02-04T17:27:19Z

A new Pull Request was created by @rgoldouz (Reza Goldouzian) for master.

It involves the following packages:

RecoEgamma/EgammaElectronAlgos (reconstruction)
RecoEgamma/EgammaPhotonProducers (reconstruction)

@jpata, @cmsbuild, @clacaputo, @slava77 can you please review it and eventually sign? Thanks.
@Sam-Harper, @jainshilpi, @valsdav, @rovere, @lgray, @sobhatta, @lecriste, @afiqaize, @wrtabb, @varuns23, @ram1123 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

slava77 · 2022-02-04T18:28:06Z

RecoEgamma/EgammaPhotonProducers/src/GEDPhotonProducer.cc

      }
    } else {
      math::XYZVector gamma_momentum = direction.unit() * scRef->energy();
      math::XYZTLorentzVectorD p4(gamma_momentum.x(), gamma_momentum.y(), gamma_momentum.z(), scRef->energy());
      newCandidate.setP4(p4);
+      newCandidate.setMass(0);


use PolarLorentzVector. This is the internal representation of reco candidate p4

Hello Slava,

The setMass function does this transformation.
https://cmssdt.cern.ch/lxr/source/DataFormats/Candidate/interface/ParticleState.h#0159

Do you mean to use,
math::PtEtaPhiMLorentzVector p4(gamma_momentum.Pt(), gamma_momentum.Eta(), gamma_momentum.Phi(), 0.0);
I could not find Pt function in math::XYZVector.

Please let me know your opinion.
Thanks,
Reza

yes, I meant to replace the XYZT p4 construction and setP4;setMass with just a construction of Polar p4 followed by setP4

Hi Slava,

As I mentioned, math::XYZVector does not have the Pt function. I need to recalculate it and I afraid to introduce bug there. What is wrong with using setMass function?
It does this transformation automatically using setCartesian function,
https://cmssdt.cern.ch/lxr/source/DataFormats/Candidate/interface/ParticleState.h#0159

Thanks,
Reza

cmssw/DataFormats/Candidate/interface/Particle.h

Line 23 in 6d2f660

typedef math::PtEtaPhiMLorentzVector PolarLorentzVector;

Sam-Harper · 2022-02-08T12:02:09Z

are we sure we want to set the electron mass to 0.5 MeV?

As far as I'm aware, all E/gamma code treats the electron as massless and has for ever (modulo this bug where it its not exactly zero due to rounding errors.

Why not continue to set it to zero as was intended. Is there is a compelling use case to need to adjust the momentum such that the mass is 0.5 MeV? It seems like such a small effect which is best ignored.

Otherwise, thanks for the fix Reza!

cmsbuild · 2022-02-08T12:44:39Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-36889/28216

This PR adds an extra 20KB to repository
There are other open Pull requests which might conflict with changes you have proposed:
- File RecoEgamma/EgammaPhotonProducers/src/GEDPhotonProducer.cc modified in PR(s): Implementation of beam halo tagger algorithm for the endcap photons #36901

cmsbuild · 2022-02-08T12:45:06Z

Pull request #36889 was updated. @jpata, @cmsbuild, @clacaputo, @slava77 can you please check and sign again.

rgoldouz · 2022-02-08T12:59:30Z

Thanks Sam,
I just updated the code with setting ele mass = 0.

Please let me know if you have any further comment.

Thanks,
Reza

cmsbuild · 2022-02-08T13:03:41Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-36889/28218

This PR adds an extra 36KB to repository
There are other open Pull requests which might conflict with changes you have proposed:
- File RecoEgamma/EgammaPhotonProducers/src/GEDPhotonProducer.cc modified in PR(s): Implementation of beam halo tagger algorithm for the endcap photons #36901

cmsbuild · 2022-02-08T13:03:59Z

Pull request #36889 was updated. @jpata, @cmsbuild, @clacaputo, @slava77 can you please check and sign again.

clacaputo · 2022-02-08T15:05:22Z

@cmsbuild please test

cmsbuild · 2022-02-08T18:54:21Z

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-36176d/22288/summary.html
COMMIT: c401a22
CMSSW: CMSSW_12_3_X_2022-02-08-1100/slc7_amd64_gcc10
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/36889/22288/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: 231 differences found in the comparisons
DQMHistoTests: Total files compared: 46
DQMHistoTests: Total histograms compared: 3765080
DQMHistoTests: Total failures: 262
DQMHistoTests: Total nulls: 1
DQMHistoTests: Total successes: 3764795
DQMHistoTests: Total skipped: 22
DQMHistoTests: Total Missing objects: 0
DQMHistoSizes: Histogram memory added: -0.004 KiB( 45 files compared)
DQMHistoSizes: changed ( 312.0 ): -0.004 KiB MessageLogger/Warnings
Checked 193 log files, 42 edm output root files, 46 DQM output files
TriggerResults: no differences found

jpata · 2022-02-09T18:33:51Z

to be a bit more verbose, I think Slava's suggestion to use more precise arithmetic through typedef math::PtEtaPhiMLorentzVector PolarLorentzVector; should be tried here.
(math::XYZVector was not suggested and seems to be a miscommunication)

rgoldouz · 2022-02-11T08:53:47Z

Hi @jpata ,
Sorry for my late reply.
As I wrote in my reply to Slava,
If I change the code like this;
math::PtEtaPhiMLorentzVector p4(gamma_momentum.Pt(), gamma_momentum.Eta(), gamma_momentum.Phi(), 0.0);
I need to use gamma_momentum.Pt(), but there is no Pt() function in gamma_momentum which is a math::XYZVector.
I think the corresponding function should be R(). But I afraid to destroy all photon object and found SetMass() function much safer

Please let me know your opinion.

Thanks,
Reza

jpata · 2022-02-11T13:23:02Z

Perhaps something like this?

math::PtEtaPhiMLorentzVector p4(gamma_momentum.rho(), gamma_momentum.eta(), gamma_momentum.phi(), 0.0);
newCandidate.setP4(p4);

ele/ph mass fix

5d1f8a1

cmsbuild added this to the CMSSW_12_3_X milestone Feb 4, 2022

cmsbuild added code-checks-pending orp-pending pending-signatures reconstruction-pending tests-pending labels Feb 4, 2022

cmsbuild added code-checks-approved and removed code-checks-pending labels Feb 4, 2022

slava77 reviewed Feb 4, 2022

View reviewed changes

cmsbuild mentioned this pull request Feb 7, 2022

Implementation of beam halo tagger algorithm for the endcap photons #36901

Merged

electron mass set to zero

ad3c65e

cmsbuild added code-checks-pending and removed code-checks-approved labels Feb 8, 2022

cmsbuild added code-checks-approved and removed code-checks-pending labels Feb 8, 2022

electron mass set to zero

c401a22

cmsbuild added code-checks-pending and removed code-checks-approved labels Feb 8, 2022

cmsbuild added code-checks-approved and removed code-checks-pending labels Feb 8, 2022

cmsbuild added tests-started and removed tests-pending labels Feb 8, 2022

cmsbuild added tests-approved and removed tests-started labels Feb 8, 2022

rgoldouz mentioned this pull request Feb 14, 2022

Electron and photon masses are set to zero #36954

Merged

rgoldouz closed this Feb 14, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ele/ph mass fix #36889

ele/ph mass fix #36889

rgoldouz commented Feb 4, 2022

cmsbuild commented Feb 4, 2022

cmsbuild commented Feb 4, 2022

slava77 Feb 4, 2022

rgoldouz Feb 8, 2022

slava77 Feb 8, 2022

rgoldouz Feb 8, 2022

jpata Feb 8, 2022

Sam-Harper commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

rgoldouz commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

clacaputo commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

jpata commented Feb 9, 2022 •

edited

Loading

rgoldouz commented Feb 11, 2022

jpata commented Feb 11, 2022

ele/ph mass fix #36889

ele/ph mass fix #36889

Conversation

rgoldouz commented Feb 4, 2022

PR description:

PR validation:

cmsbuild commented Feb 4, 2022

cmsbuild commented Feb 4, 2022

slava77 Feb 4, 2022

Choose a reason for hiding this comment

rgoldouz Feb 8, 2022

Choose a reason for hiding this comment

slava77 Feb 8, 2022

Choose a reason for hiding this comment

rgoldouz Feb 8, 2022

Choose a reason for hiding this comment

jpata Feb 8, 2022

Choose a reason for hiding this comment

Sam-Harper commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

rgoldouz commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

clacaputo commented Feb 8, 2022

cmsbuild commented Feb 8, 2022

Comparison Summary

jpata commented Feb 9, 2022 • edited Loading

rgoldouz commented Feb 11, 2022

jpata commented Feb 11, 2022

jpata commented Feb 9, 2022 •

edited

Loading