Remove iterativeConeGenJets from CMSSW, Add GenJetsNoNu to RECO/AOD

[ahinzmann]

Two long outstanding issues addressed in this pull request:

• Remove iterativeCone6GenJets from RECO+HLT
• Prepare switch from ak4GenJets to ak4GenJetsNoNu in RECO+AOD
  Discussed at JMAR meeting:
  https://indico.cern.ch/event/357182/contribution/5/material/slides/0.pdf
  After a more general presentation, to discuss the change, but before the last 7_4_X pre-release, we would like to remove the ak4GenJets collection, together with the other jet collections discussed at the RECO meeting today.
  https://indico.cern.ch/event/354427/session/4/contribution/8/material/slides/0.pdf

Details:

• Add ak4GenJetsNoNu and ak8GenJetsNoNu and remove iterativeCone5GenJets from RECO+AOD
• HLT jet validation use now ak4GenJets instead of iterativeCone5GenJets as truth reference.
• Remove genMetIC5GenJets, genMetCaloAndNonPrompt.
• PAT is not yet touched and ak4GenJets are kept for a pre-release cycle because it needs new RelVals before changes are possible and ak4GenJets can be dropped

[cmsbuild]

A new Pull Request was created by @ahinzmann for CMSSW_7_4_X.

Remove iterativeConeGenJets from CMSSW, Switch from GenJets to GenJetsNoNu

It involves the following packages:

DQMOffline/RecoB
FastSimulation/HighLevelTrigger
FastSimulation/Validation
HLTrigger/HLTanalyzers
HLTriggerOffline/Higgs
HLTriggerOffline/JetMET
HLTriggerOffline/SUSYBSM
JetMETCorrections/MCJet
RecoEgamma/Examples
RecoJets/Configuration
RecoMET/Configuration
RecoMET/METProducers
RecoTauTag/TauTagTools
Validation/EventGenerator
Validation/RecoB
Validation/RecoEgamma
Validation/RecoJets
Validation/RecoParticleFlow
Validation/RecoTau

@vciulli, @perrotta, @civanch, @covarell, @bendavid, @lveldere, @danduggan, @fwyzard, @thuer, @monttj, @Martin-Grunewald, @cmsbuild, @nclopezo, @mdhildreth, @deguio, @slava77, @vadler, @StoyanStoynev can you please review it and eventually sign? Thanks.
@rappoccio, @Sam-Harper, @imarches, @acaudron, @jpavel, @TaiSakuma, @lgray, @inugent, @nhanvtran, @mmarionncern, @schoef, @ferencek, @matt-komm, @mariadalfonso, @pvmulder, @richard-cms, @rociovilar this is something you requested to watch as well.
You can sign-off by replying to this message having '+1' in the first line of your reply.
You can reject by replying to this message having '-1' in the first line of your reply.

[bendavid]

Few comments.

1. I think the change of the default gen jet collections needs to be more widely discussed.

2. GenJets are run at GEN-SIM time, so any change to the default genjet sequence is too late for the run 2 production. (Unless we run the modified sequence to re-run the genparticles/genjets at DIGI-RECO time for all future DIGI-RECO campaigns)

3. GenJets can clearly be run on the fly at analysis level (or at MINIAOD-time)

[ahinzmann]

Hi Josh,

1. Right, the change can only happen after the general JetMET talk (probably in January). But in order to have a chance to make the change, both collections already need to be in a pre-release before that, because otherwise PAT validation based on RelVal will not work. I suggest that I modify the request, such that it only adds the new GenJetNoNu collection and drops the iterativeCone, while keeping everything else as is. Then this is just a preparation step for the actual change (if decided in a general meeting).
2. Ok, so it's already too late:-) But maybe it is still a good idea to make the change now, otherwise we'll just never do it in the future and always use reclustered GenJets as for many years now.
3. Sure. But if we manage to have the right collection in RECO/AOD (in the future) we can save computing time at the analysis level.

[slava77]

+1

for #6884 7c5a638

tested on higher stats in
CMSSW_7_4_X_2014-12-18-0200 (test area sign477).
all tested workflows recycled the gensim input from an older pre-release.

• There are no changes in monitored RECO quantities at direct object comparison;
• There are quite a few changes in the DQM outputs all coupled to the change in the reference jets used for denominator or matching. These changes are generally mild (attached below)

The additional gen jet collections added in RecoGenJets_cff will show up only after GEN-SIM inputs are remade.
Since 71X GEN-SIM is going to be used for MC production, the production DQM sequences will not have these new gen jets monitored unless they are remade on the fly.

flat-qcd spectrum 15 GeV to 3 TeV
just a few examples of changes in distributions