Readiness for production

rete commented 6 years ago

@gaede @shaojunlu @danieljeans I post this issue as a central post for the things we need to check before going to the final production. Here is the proposal list:

[x] JER/JES (uds)
[x] Particle ID (ttbar and single particles)
[x] dEdX (ttbar and single particles)
[x] Tracking performances (ttbar and single muons)
[x] Energy calibration (kaon0L, muons, photons)
[x] MCInfo such parent-daughter relations, etc ... (one physics channel, e.g 1 TeV WW/ZZ->qqqq)
[x] Background distributions (pure background simulations)
[x] Event numbering after DST merge
[x] LCFIPlus vertexing and flavour tagging

Please, post a comment or edit this comment to add a new entry or validate one entry as soon as we have the plots to validate it. Post also the plots to keep track of it.

shaojunlu commented 6 years ago

@rete I saw that you have the pure background simulations in the list. Could I use them in standard reconstruction to check the impact on the tracking efficiency? I may copy the ttbar simulation/reconstruction files into the ILD group folder, too.

rete commented 6 years ago

@shaojunlu I suggest we just check the tracking and jer/jes for the pre-release(s). For the release v01-19-06 the central production will be run and we will then go through this list.

rete commented 6 years ago

PID performance

Dear all,

I have run the PID scripts available in the ILDPerformance package. I ran this on ttbar events (REC files) centrally produced from the 3rd test production available from :

/ilc/prod/ilc/mc-opt-2/ild/rec/500-TDR_ws/6f_ttbar/ILD_l5_o1_v02/v01-19-06-p01/00009663/000/

After running the PID performance scripts I get the plots available there :

http://www.desy.de/~eteremi/3rd-test-prod/PID/

Note that the plots (PIDs_pdg*.pdf) are produced by selecting particles with momentum > 3 GeV.

I found that :

the dEdX plot (input to the likelihood PID) seems reasonable compared to what was presented by U. Einhaus a few weeks ago
the muon PID efficiency is really low as estimated by the final likelihood algorithm. In particular the efficiency seems really nice as estimated by PandoraPFA (BasicPID) and by the shower shape algorithms. For p > 3 GeV, the dEdX PID estimation starts to be bad but it seems to have a huge impact on the final PID...
same comments for the electrons efficiency as for muons: the efficiency is better with PandoraPFA and shower shape but worse for the final likelihood
the dEdX PID estimation for pions seems weird to me. From the dEdX plot I understand that pions and muons are almost mixed and can't be separated in the region p > 3 GeV. I would have expected to have a pion PID with something like half muons / half pions which is not the case.

rete commented 6 years ago

JER/JES performance

Dear all, this is the performance plots from ILDPerformance run on uds samples for iLCSoft v01-19-06

Jet energy resolution
Jet energy scale and deviation from linearity

This has already been shown by Frank during the last ILD phone meeting.

danieljeans commented 6 years ago

maps of simhits produced by muons uploaded to confluence: https://confluence.desy.de/display/ILD/Simulation#Simulation-Validationofsimulation

no obvious drastic problems seen

rete commented 6 years ago

Hi Daniel,

maps of simhits produced by muons uploaded to confluence: https://confluence.desy.de/display/ILD/Simulation#Simulation-Validationofsimulation

it seems the plots can't be seen on page. Even by clicking on them, the pictures are wrong.

danieljeans commented 6 years ago

when I click on it, I can download a ps file full of plots, as intended.

it's true that confluence doesn't seem to preview ps files.

danieljeans commented 6 years ago

detector material scans for latest version at http://research.kek.jp/people/jeans/ctpics/

no obvious problems noticed

rete commented 6 years ago

Calibration

Current calibration from v01-19-05, currently in the software :

<constant name="EcalBarrelMip">0.0001525</constant>
<constant name="EcalEndcapMip">0.0001525</constant>
<constant name="EcalRingMip">0.0001525</constant>
<constant name="HcalBarrelMip">0.0004925</constant>
<constant name="HcalEndcapMip">0.0004775</constant>
<constant name="HcalRingMip">0.0004875</constant>
<constant name="EcalBarrelEnergyFactors">0.00616736103247 0.0125274552256</constant>
<constant name="EcalEndcapEnergyFactors">0.0064868449976 0.0131764071919</constant>
<constant name="EcalRingEnergyFactors">0.0064868449976 0.0131764071919</constant>
<constant name="HcalBarrelEnergyFactors">0.0216747245411</constant>
<constant name="HcalEndcapEnergyFactors">0.0217395864899</constant>
<constant name="HcalRingEnergyFactors">0.0271318181372</constant>
<constant name="MuonCalibration">56.7</constant>
<constant name="PandoraEcalToMip">153.846</constant>
<constant name="PandoraHcalToMip">43.29</constant>
<constant name="PandoraMuonToMip">10.3093</constant>
<constant name="PandoraEcalToEMScale">1.0</constant>
<constant name="PandoraHcalToEMScale">1.0</constant>
<constant name="PandoraEcalToHadBarrelScale">1.07522318318</constant>
<constant name="PandoraEcalToHadEndcapScale">1.07522318318</constant>
<constant name="PandoraHcalToHadScale">1.02821419758</constant>
<constant name="PandoraSoftwareCompensationWeights">1.66803 -0.031982 0.000192898 -0.0612971 0.00256256 -4.35641e-05 0.0558589 0.0601767 -0.0758029</constant>

New calibration obtained with v01-19-06 :

<constant name="EcalBarrelMip">0.0001575</constant>
<constant name="EcalEndcapMip">0.0001575</constant>
<constant name="EcalRingMip">0.0001575</constant>
<constant name="HcalBarrelMip">0.0004975</constant>
<constant name="HcalEndcapMip">0.0004825</constant>
<constant name="HcalRingMip">0.0004875</constant>
<constant name="EcalBarrelEnergyFactors">0.00637340690428 0.0129459860072</constant>
<constant name="EcalEndcapEnergyFactors">0.00670894460888 0.01362754714</constant>
<constant name="EcalRingEnergyFactors">0.00670894460888 0.01362754714</constant>
<constant name="HcalBarrelEnergyFactors">0.0216747245411</constant>
<constant name="HcalEndcapEnergyFactors">0.0217395864899</constant>
<constant name="HcalRingEnergyFactors">0.0270819056524</constant>
<constant name="MuonCalibration">56.7</constant>
<constant name="PandoraEcalToMip">153.846</constant>
<constant name="PandoraHcalToMip">43.29</constant>
<constant name="PandoraMuonToMip">10.3093</constant>
<constant name="PandoraEcalToEMScale">1.0</constant>
<constant name="PandoraHcalToEMScale">1.0</constant>
<constant name="PandoraEcalToHadBarrelScale">1.05666809149</constant>
<constant name="PandoraEcalToHadEndcapScale">1.05666809149</constant>
<constant name="PandoraHcalToHadScale">1.06268778947</constant>
<constant name="PandoraSoftwareCompensationWeights">1.45867 -0.0137705 8.64291e-06 -0.0492179 0.00128169 -1.44578e-05 0.054839 0.0539599 -0\
.0474883</constant>

Single neutral hadron energy before and after software compensation : softwarecompensationmonitoring_totalenergy

The results are consistent by seeing the plots. Nevertheless, the calibration constants are different enough to submit a pull request in ILDConfig.

@gaede , FYI

rete commented 6 years ago

Everything has been checked so far. I close the issue. Thank you all !

iLCSoft / ILDConfig