Coming Soon


Nov 26: Tuesday

YM and Roger Rusack wanted me to see if I could do fluka simulations for PFCAL, now what is this exactly. Attended HGCAL ( ) meeting. There is a lot this group has to do before they have final configuration for me - it looks like to me - I would ask Roger/Yurii. Anyways spent full 9-12 hours for this meeting. I'll see what is there in their twiki:

Now its time to leave for Sadia.

Nov 25: Monday

No group meeting - bosses are busy

Edited ERecoNote with references and spelling mistakes - Ilya would send comments with in 2-3 days.

Tom sent email for SFs 2-3 days ago about difference in accessing the GSF reference from PAT- even to Ilya - but it looks to me the problem he mentioned is only because of different methods used for eta - I am using SC-eta and Gsf-Pt for SF calculations.

Alfredo want to know what exactly I did for upgrading CFC geometry - sent email - no time for skype - so sent my geometry files to him.

Ilya need event list for TnP -data- which he could compare to see how much pass-fail count differ for us. Here is a event list from data: for events with - mass 60 -120 - probe_sc_abseta 1.566-2.0 - probe_gsfEle_pt 15-20. Print out format is: " run event mass tag_gsfEle_pt tag_sc_abseta probe_gsfEle_pt probe_sc_abseta probe_isWPMedium" .

What Yurii uses for estar -sent emails and final is: SCeta and ElePt - I am not what happen to the concept of bad tracker recons. for high Pt elecs. ( think more) Note : Et= energy*sin(theta) or energy/cosh(eta)

Nov 22: Friday

Estar again: Here is long list of plots but for M=1000GeV and with one of the background sample: DYJets. I am doing shape comparison to see where could be possible cuts for Tau's and other invariant mass variables. I tried to follow AN-2013/263 ( dated: Nov 8, 2013) - there it is mentioned that I should cut on supercluster Et instead of Pt, max vs min mass plot is not very clean ( index.html#MassElec1Ca8Jet_vs_Elec2Ca8Jet_invMassCut.png ). cuts on tau's to reject backgrounds would help here - explore more.

ToDo - skim the samples as per my need - do proper weighting according to lumi and cross-section.

Nov 21: Thu

How to do PURW for ggntuples? emailed Misha - he replied - using - PU distribution for data from Ming - /uscms/home/makouski/public/Pileup_Observed_69300.root

so done. setting done according to my analyzer.

Nov 20: Wed

Nov 19: Tue

CFC envelope: from Nural's file - dimensions for CFC envelope along z are 317-520cm, which is OK, I did that. But assuming a PE-absorber layer of 25cm before CFC-block need z=292-317cm. But in present CMS geometry, I find tracker-avergae material upto 297.3cm. Is it OK to play with end dimensions of tracker? Should I decrease tracker volume length along Z or decrease the PE width to 20cm?

Sapta need TnP docs for her thesis.

Nov 18: Mon

For Yurii: he asked for quick estimate of efficiency as a function of pT for barrel and endcap for the reconstruction of electrons (just MC) for H/E and sieie cuts. I am done with this - using WToENu from pythia: "/WToENu_TuneZ2star_8TeV_pythia6/Summer12_DR53X-PU25bx25_START53_V19D-v1/AODSIM" Selection done for:

Nov 13: Wednesday

I have done a quick analysis for m800 and m400 samples. Shapes are compared here: - Along with above two samples produced by YM, 3rd one with m=1000 was from Ming with 10K events.

Nov 6: Wednesday

Merge all the samples - data as well as mc. I am using dyjets ( ) which have 3503b as cross section ( ), this has 30459503 events, so its weight : 19.6*3503710/30459503 = 2.25.

Nov 5: Tuesday

Will work on adding tnp producer today. Wow, I am done with tnp producer. OK done, with lot of effort on coding such producer, I have all these variables ready to put on TnPtree. But still this valuemap is not accepting any other type otherthan float. why? I will see later, for the time being float thing would work. I need to submit crab jobs on data because wisc people deleted my samples from their computers. Done, crab jobs are submitted and condor for mc sample.

Nov 4: Monday

Group meeting done.

myTnP tools are really in bad shape. only I can deal with that. But before cleaning that stuff, I want to add more stuff like add in tree variables for ID. but what about isolation and conversions. this is something which is not coming from electron object directly. need to add producer with value maps for d0vtx, dzvtx, isolations, vtxFitconversion.

Nov 1: Friday

preparing slides for fcal meeting to be at 10AM. fluka cfc now brass: ( brass doing great job, less neutron flux. SKunori asked for full dose map and send that txt file to sasha. Done.

come back to TnP stuff. I need to spend some time on my tools, before I pass this to github for Chiara n all.

Oct 31: Thursday

I'll be at home, not well

Why FLUKA is not working on my dell...hurrr it says this fluka version is not supported anymore. need to move to latest fluka. I would do this on mac now.

mac don't make this fluka to compile. tried all g77/gfortran versions. fluka 64, 32 bit nothing is working for this mac. I hate this mac. wasted full day ( its 3PM already). need to move back to dell ( my laptop will die soon because of this fluka).

wow, finally fluka2011.2b-linuxAA.tar.gz working on my dell now. good !!!

make brass flux and dose maps. compare to nominal cms as well as with W-CFC. After spending 4 hours, finally concluded that brass has less damage as compared to tungsten. Will prepare slides in morning.

Oct 30: Wednesday

will work from home

wgg: ( we seem to have potential tight limits as compared to existing theory ones) Talk and all finished by 11.30AM. (sneezing+cold- need rest, can't see screen)

Nural askd for fluka talk: I'll do this tomorrow.

Oct 29: Tuesday

YM: Give a talk on wgg tomorrow. Now leave this TnP stuff and read those wgg docs. What I have done for SM vs aGCs: , why only C_www and CP_www sensitive not others: YM says "they control different terms that might be less strongly contributing to the overall amplitude, so the sensitivity might be lower." but if shapes are different as we see in case of above two - then we need to produce lots of MC samples for different couplings. IS has done all limit setting thing. Produce Ptgamma plot with variable binning (where last one is 200GeV and above) from SM as well as aGCs ( only sensitive one: C_www and CP_www). Scale plot to 20fb^-1 and assuming some elec/photon efficiencies. Generate pseudo data from SM using Poisson statistics. Plot this last bin yield Vs coupling values ( look like parabola), fit with parabola and generate many aGC points. Take LLH using d0 method-chi2statistics for each point and plot as a fn of coupling values. Fit with parabola and see what we get for 95%CL.

Oct 28: Monday


webpage done: edited twiki : ..want to plot all shape variables. It seems like I will spend lot of time on this.

Oct 25: Friday

no cfc - US cms meeting - ndhngra -pu

Need to run fluka on mac...compiler issue...but my dell is dieng ... hurrr !

prepare some slides for egamma on monday

Oct 24: Thursday

YM meeting done

IS explained his LLH stuff.

1MeVneq added to slides

Oct 23:

YM meeting done

IS: running wgg mg5 jobs for a number of different values for cwww - the main coupling showing sensitivity in pt-photon

Need to plot fluka ratios in tracker and preshower region (up to 322 cm) for Milena. She need a ratio w.r.t original CMS geometry ( I need to find those maps from the number of simulations I performed during last 4-5 months frown ). her talk for tomorrow: I modified CMS geometry from Z=297.3 to 322.3cm where originally was preshower + some absorbing materials. Actually preshower is (, ) is from Z=298.5 - 316.5cm So, finally I have ratio plots:

  • slide6: shows the fluence maps from Nominal CMS geometry ( don't get confused by geometry overlay - which is modified for CFC - I did that intentionally to observe the trends w.r.t in new CFC detector)
  • slide7: is a compariosn of neutron/photon fluences in CFC w.r.t Nominal - which is less than 2 times for tracker.

time for estar: Ming sent me link to pubslished data: the sample was published in TWT2. the dataset path is /pythia_estar_eejj_L10000_m1000_GEN_SIM_v2/cmkuo-pythia_estar_eejj_L10000_m1000_RECO_v2-3664d28163503ca8171ba37083c39fc9/USER, looking at slides:, I have: , why such shapes...never seen before frown

Oct 22:

YM meeting done

IK agreed on my point: he has ABSOLUTE systematic error on the efficiencies, not on the scale factors. Now his method would be " one should take the absolute error from each source and divide it by either data or MC efficiency (which most of the time are relatively close) to get the relative error from this source." He has a list of what efficiency to use as the denominator for individual sources:

   fit model: from data
   fit binning: from data
   tag dependence: from data
   pile-up: from MC
   generator dependence: from data
he sent the new file, now see this example: for the bin of choice, 10-15 GeV and 0.0-0.8 for eta, one finds there 11.0%, which means that the error on the scale factor for that bin is 0.842 +- 0.018(stat) +- 0.093 (syst).

I could not get slot in EGamma meeting, so it would be on coming Monday.

I spend 2 more hours on this...still confused, but IK helped me to understand his way of evaluating all this...good now, will prepare slides.

Milena from CERN asked for FLUKA plots that I shoed in workshop. I made the ratio stuff and here all sent to her:

Oct 21: Monday

group meeting done

will send MG5 to irakli : there are many things to add to Madg5 twiki: done

ScaleFactor: need to remake templates for loose/veto/medium---forget to add tau-removal thing frown but there is something wrong with ilya's systematics: I am looking at these talks: ( I would call it main) and ( updated ).

from main

  • slide-14 gives a table with data and mc efficiencies and SFs for medium ID.
  • slide 23: gives systematic due to choice of model to fit.

from updated

  • slide 27 says: table is for updated systematics on eID SFs.

let us have a look at first bin with pt(10-15) and eta(0.0-0.8):

Data effi.: 0.379
MC effi.:   0.450
SF:       : 0.842

systematic errors:
fit model: 3.6
fit binning:  0.53
tag dependence:  1.1
pileup:        o.73
gen-level : 1.7
max. systematic : 4.3

but now to obtain the systematic on the scale factors I would expect to divide the number above by the efficiency on MC. Consider one example say with fit model numbers. If you take CB fit model which is responsible for maximum error in this bin (fit model systematic error: 37.54-33.99 = 3.6), gives you scale factor 33.99/45.0=0.755. So, SFs have a difference of 0.086, either 8.6% directly or 0.086/0.842 = 0.103 means 10.3% relative/fractional systematic error should have been quoted as per my understanding.

Oct 19: Saturday

fluka jobs submitted for 1MeVneq but failed because of disk quota exceeded again

Edited CMS geometry to incorporate this new detector - brass block of 200cm along z ... it take do the three set of jobs, no PE, with PE, PEdoped

literature survey : systematics:

What do I need to do for SFs: systematics and evaluate effi. for data and mc for other three loose/veto/tight IDs: My last presentation was for medium id only where I figurd out main few differences with Ilya, but ofcourse that doesn't include isolation and all because this was found after this talk. Now for summary he impose veto on charged hadrons which share same ctf-track as that of electron and ,

Oct 18: Friday

FLUKA day: shown updated maps in meeting:, sasha did his part with new dose map, all set now as per expectations. ToDO try to do simulations when there is a brass block instead of W but size of block along Z should be 2m - Nural + Phil need this. Also score 1MeVneq which would help to understand

About FLUKA DOSE routine: : Conversion coffiecients for converting fluence into dose for incident neutrons, protons, photons, electrons, muons and charged pions

Neutron fluence: ,

USCMS meeting:

YM introducing irakli

sudha helped for x11 in mac: Download links - use X11 2.6.3 - 2011.07.20 - Final Leopard Release

I again wasted day with this mac - now FLUKA is not installing because of some compiler issue.

W&C Seminar tomorrow at 4pm CDT - Kai Yi (Iowa) on "Observation of a Structure in J/ψφ from B+→ J/ψφK+ Decay at CMS"

Oct 17: Thursday

I have some scripts ready (but not so good) for unzipping the lhe files, merging all the lhe files and then converting them to root format. But since I have meeting in next 2 hours ( at 7AM UStime), so I need to just make slides with whatever I have already i.e. from C(www)@16 Vs SM. Here I am done with this: . Lindsey: Matt: Yurii:

Oct 16: Wednesday

FLUKA: Moritz group discussed my slides: ,

Could not attend WGM : -- will see the slides

Yurii quick chat: shared my wgg stuff from last night: . need to use variable binning to see excess in tails. Will use big number for various couplings like 16 or ss, as lindsey+yurii used in their talk in dec2012.

I am again wasting time on root and mac, trying to use root 5.28 this time. Sadia/Sudha I need X11 to get installed again

oh no cvs dead today: need to move to git.
git init
git add
git commit -m "first commit"
git remote add origin
git push -u origin master
git remote add origin
git push -u origin master

cern is at slc6 now

  cd .ssh/
  ssh-keygen -t rsa -C ""
  sudo apt-get install xclip
  xclip -sel clip < ~/.ssh/
 ssh -T

cmsrel CMSSW_6_2_0_pre7
cd CMSSW_6_2_0_pre7/
cd src/
cvs2git --blobfile=git-blob.dat --dumpfile=git-dump.dat /afs/ --symbol-transform="(.*)/:\1-" --use-external-blob-generator --fallback-encoding "UTF8" --username Lovedeep
git push origin master
set MY_GITHUB_USER `git config --get user.github`
setenv MY_GITHUB_USER `git config --get user.github`
setenv MY_REMOTE`git config --get user.github`/UserCode.git

git init UserCode
cd UserCode
git remote add origin $MY_REMOTE
cat ../git-blob.dat ../git-dump.dat | git fast-import
git config http.postBuffer 524288000
git push --mirror -u origin

Need to merge all lhe files on cmslpc - i generated around 2000 files. First need to do gunzip for these lhe files in eos. A problem file has 1 other link -- unchanged , OK use -f as mentioned here ( )

Oct 15: Tuesday

Done with Ersh doc - 9.15AM

want to read about US shutdown impact on fnal:

Sophie from Fluka want me to show up in - - Radiation Simulation General Working Group Meeting #7 - Tuesday, 15 October 2013

Root on mac not working: why? Everything installed properly: before entering root add gfortran and gcc from

mv ~/Downloads/gcc-4.8-bin.tar.gz .
gunzip gcc-4.8-bin.tar.gz
sudo tar xvf gcc-4.8-bin.tar

same did for gfortran-4.8-bin.tar

Enabled support for asimage, astiff, bonjour, builtin_afterimage, builtin_ftgl, builtin_glew, builtin_pcre, builtin_lzma, cintex, explicitlink, fink, genvector, krb5, ldap, memstat, odbc, opengl, python, reflex, rpath, shared, sqlite, ssl, tmva, x11, xft, xml.

make -j 4

   ===                ROOT BUILD SUCCESSFUL.                ===
   === Run 'source bin/thisroot.[c]sh' before starting ROOT ===

source bin/

root : this stay in hanging situation , no splash , no display...but if root was installed properly, it should have given a mesage like: *** DISPLAY not set, setting it to localhost:0.0

may be I was using root 5.34, let me try with 5.28.

MG5: about randm seed in cards: I changed iSeed parameter in run_card.dat, now it is added externally and would be different for every job. Note: just checked donot generate more than 100K events per run, plot lead pt gamma, mt for wgg mainly, increase stat, run jobs in parallel at lpc.

Oct 14: Monday

Group meeting by Andrew - done

Our paper SMP-12-004 has been appeared on the arXiv today: smile

Got explanations for Moritz query: actually after doing the upgrades to geometry, I submitted three different jobs for simulations. And somehow in one of the job where I added PE shielding, I didn't compiled the user routines which means magnetic field was not in. Now when I do the simulations with proper magnetic field and linking dpmjet3, I get everything as per the expectations. See my slides here: Polyethylene is doing its job perfectly and B-10 doped PE is even better for neutron shielding. I think I should update the workshop slides to reflect the proper maps there.

Reading papers for MG5 - ewdim6 model, Also will try to get some results for this.

Oct 10 and 11: Thursday, Friday ( Photons- Hands-on Tutorial Session (HATS) at LPC (HATS@LPC)): very interesting exercises on clustering algorithms - did on paper- it was very interesting - Andrew's idea of doing the exercises on paper was excellent smile - I learnt many things smile

LPC news: Fermilab shutdown, everyone listening to news frown

Oct 9: wednesday

Ershu sick bcz of antibiotics n side-effects frown

At 4AM: joined vidyo for rad-sim fluka workshop, I was supposed to talk at 5.20, but could do it around 6.20AM: Why B10-PE has more neutron fluence than PE only - Moritz. Mika thesis: Although Boron is one of the important material for thermal neutron shielding, but if we have high neutron energies then boron has some inelastic reaction channels which can lead to an increase in neutron fluences.

The approved calendar for the CMS Weeks and events for 2014 is available at

WGM meeting : - interesting news

To Ilya: He has doubts on isolation calculation in my ntuples ( which i also had in past wink ). see this:

> event no. 10591606 .
> [ PFCandidate: Type: 4  et:  23.9142  Pt:  23.9142  Eta: -0.376352  Phi: -1.66],   [  delR: 0.0267107]
> [ PFCandidate: Type: 1  et:  12.9892  Pt:  12.9885  Eta: -0.375337  Phi: -1.687],   [  delR: 0.000337959]
> SumPtCharHad :12.9885
As per him, 12.989 GeV particle should fail "footprint removal" requirements ( ) - under "recommended electron vetos (april 4 2012"). but this is in barrel - he ignored this point I guess. Another thing he askd: "How do you decide whether to drop a charged hadron or not from isolation sum if it matches your electron? (Do you have check like that at all?)". I donot know the exact answer, either we can ask this from others on hypernews or I will try to find sometime to figure it out today.

wow Moritz happy with my work smile - To DO : need time to answer his email.

Oct 8: Tuesday

To Do:

  • a pizza party this Tuesday the 8th at 6pm in the User Center Music Room to announce and celebrate with the new officers
  • read IK email for isolation etc.

Nobel prize - Higgs

From LPC: , )

YM sent emails to l* group: we would be ready with Z-jj analysis within 1 month.

Sasha got everything he need from fluka dose maps. results match with 2012 data. He presented his method here : - develop a method for estimation of photon efficiency and time of travel inside cfc fibers, update on radiation damage model for cfc, -dose vs eta plots, - dose vs eta with my cfc dose map - something wrong? add without pe layer - he did this today, so now everything seem ok :

131008 01.png

prepared for Fluka workshop talk.

Oct 7:

KSU group meeting: discussion on ggNupls, YM "about to discover" new physics ... YM would be in US in Oct end...gothrough QGC talks...

I have re-run the MG5 generation using EWdim6 model with only one parameter "CWWWL2" as 1 while the rest 4 parameters as 0. Last time all of five dim6 parameters were set to 10. I am plotting invariant mass of Waa system, transverse mass of W , leading photon transverse momentum - as suggested by YM during meeting - along with some other variables. This time I used cmslpc condor for submitting a job with 20K events for dim6 model. I actually submitted single job because I could not find a quick way to change random seed for submitted jobs in parallel. Otherwise scripts are all set for submitting jobs there. I tried to find Patricia during the day, but she was not in office today. I sent her email also, but no response.

Oct 6

YM emailed once "I wanted to check if excited lepton with ell* -> Z e -> jj e would be a much better choice than Z->ell ell channel" - so I added merged jets in the ntuple. pp->eestar, estar->eZ, Z->jj or merged jet. signal final state is 2e 2j, see this: ,

ggntpls store so many jet variables ...why? see this pileupjetid page: , , Do I need Pileup Jet ID for ca8 jets? we are not very sensitive to these jets because of high pt thresholds for our lepton star or Wgg analyses. But ggntuplizer, ( line number: 4625 to 4803 ) has more than 20 branches for pileup jets ID variables, along with three pre-calculated flags for Pileup jet ID. These three flags are stored as per jet-met prescriptions ( ) and can be used as such without ever having to use those 20 additional branches, which seem like doing nothing other than occupying space if we are not going to perform extensive studies dedicated towards understanding pileup jet contributions.

Oct 5
00076         if (pfCandidate->particleId() == 1) ChargedHadEt += et;
00077         if (pfCandidate->particleId() == 2) ChargedEMEt += et;
00078         if (pfCandidate->particleId() == 3) MuonEt += et;
00079         if (pfCandidate->particleId() == 4) NeutralEMEt += et;
00080         if (pfCandidate->particleId() == 5) NeutralHadEt += et;
00081         if (pfCandidate->particleId() == 6) type6Et += et;
00082         if (pfCandidate->particleId() == 7) type7Et += et;

for Ilya: the list of candidates from "pfNoPileUp" collection, who enter the calculations for charged hadron component of the pfIso deposit for electron in:

the event no. 10591606 .
Condition:   delR (electron, pfCandidate) < 0.3
electron number: 1
[ PFCandidate: Type: 4  et:  23.9142  Pt:  23.9142  Eta: -0.376352  Phi: -1.66],   [  delR: 0.0267107]
[ PFCandidate: Type: 1  et:  12.9892  Pt:  12.9885  Eta: -0.375337  Phi: -1.687],   [  delR: 0.000337959]
[ PFCandidate: Type: 4  et:  1.25753  Pt:  1.25753  Eta: -0.355614  Phi: -1.9054],   [  delR: 0.219593]
[ PFCandidate: Type: 4  et:  0.899591  Pt:  0.899591  Eta: -0.358294  Phi: -1.95042],   [  delR: 0.264277]
[ PFCandidate: Type: 4  et:  0.544959  Pt:  0.544959  Eta: -0.504453  Phi: -1.5646],   [  delR: 0.177796]
[ PFCandidate: Type: 4  et:  0.612495  Pt:  0.612495  Eta: -0.234182  Phi: -1.51999],   [  delR: 0.218346]
[ PFCandidate: Type: 4  et:  0.626602  Pt:  0.626602  Eta: -0.558515  Phi: -1.51454],   [  delR: 0.251468]

SumPtCharHad :12.9885
SumPtCharEM :0
SumPtMuon :0
SumPtNtEM :27.8553
SumPtNtHad :0

event: 2128449
Condition:   delR (electron, pfCandidate) < 0.3
electron number: 1
[ PFCandidate: Type: 2  et:  31.8494  Pt:  31.8494  Eta: 0.35532  Phi: -0.282385],   [  delR: 9.10998e-05]
[ PFCandidate: Type: 1  et:  1.48987  Pt:  1.48409  Eta: 0.357674  Phi: -0.0995274],   [  delR: 0.182782]
[ PFCandidate: Type: 1  et:  0.819216  Pt:  0.808313  Eta: 0.307739  Phi: -0.261781],   [  delR: 0.0518129]
[ PFCandidate: Type: 4  et:  0.394966  Pt:  0.394966  Eta: 0.527203  Phi: -0.130366],   [  delR: 0.229405]
[ PFCandidate: Type: 4  et:  0.59138  Pt:  0.59138  Eta: 0.202056  Phi: -0.0610419],   [  delR: 0.26915]
SumPtCharHad :2.2924
SumPtCharEM :31.8494
SumPtMuon :0
SumPtNtEM :0.986346
SumPtNtHad :0

Condition:   delR(electron, pfCandidate)<0.3
electron number: 2
[ PFCandidate: Type: 2  et:  20.0486  Pt:  20.0486  Eta: -1.72067  Phi: -2.8073],   [  delR: 0.00118305]
[ PFCandidate: Type: 1  et:  3.85529  Pt:  3.85499  Eta: -1.7236  Phi: -2.86299],   [  delR: 0.054585]
[ PFCandidate: Type: 4  et:  0.717439  Pt:  0.717439  Eta: -1.81781  Phi: -2.59434],   [  delR: 0.235143]
SumPtCharHad :3.85499
SumPtCharEM :20.0486
SumPtMuon :0
SumPtNtEM :0.717439
SumPtNtHad :0

Oct 4:

Ersh nt well

FCAL CFC meeting: I added/removed/dopped polythene in front of W-HF ( added dose map for polythene doped with B-10...It took a lot of time for me to learn to do such doping thing for fluka tool, it was just last night that I come up with something and got the results for this new material.) tables and maps for all three setups i.e. CFC with polythene, CFC without PE, CFC with B10-PE:

Tables attached to this link: First colmn is Z position, then we have r in 2nd, and dose rates in 3rd column - and dose rates are scaled to 100/fb.

TODO: Anne: , I need to prepare for a tutorial/talk for this workshop: "FLUKA simulations with W based CFC for CMS upgrades"

to IK: I am printing isolation components from following file ( DYJets-Madgraph ): /store/mc/Summer12_DR53X/DYJetsToLL_M-50_TuneZ2Star_8TeV-madgraph-tarball/AODSIM/PU_S10_START53_V7A-v1/0001/6C495815-BCD3-E111-92A0-0025B3E05CC0.root

Here is a print from first few events of this file :

Run: 1 , lumi: 26501 , event: 10591606 has electron with pt: 23.368423 , eta: -0.378367, phi: -1.686688 , pass medium(0), pass tight(0)
isolation is: 1.64598 AEff, rhoPrime, iso_nh, iso_em (iso_n), iso_ch, iso:  0.13        18.3073 0       27.8553 ( 25.4754 ) 12.9885     1.64598

Run: 1 , lumi: 26501 , event: 10591613 has electron with pt: 4.221165 , eta: 0.153634, phi: -2.447897 , pass medium(0), pass tight(0)
isolation is: 0.140728 AEff, rhoPrime, iso_nh, iso_em (iso_n), iso_ch, iso:  0.13       16.7619 0       2.77308 ( 0.594034 ) 0  0.140728

Run: 1 , lumi: 26501 , event: 10591610 has electron with pt: 29.197144 , eta: -0.028526, phi: 1.365838 , pass medium(1), pass tight(1)
isolation is: 0.0337572 AEff, rhoPrime, iso_nh, iso_em (iso_n), iso_ch, iso:  0.13      9.62613 0       0.418001 ( 0 ) 0.985613 0.0337572

Run: 1 , lumi: 26501 , event: 10591615 has electron with pt: 5.373703 , eta: -2.310954, phi: 0.280475 , pass medium(0), pass tight(0)
isolation is: 0.260269 AEff, rhoPrime, iso_nh, iso_em (iso_n), iso_ch, iso:  0.11       28.8466 0       1.13015 ( 0 ) 1.39861   0.260269

Run: 1 , lumi: 26501 , event: 10591621 has electron with pt: 4.191663 , eta: 1.954083, phi: 0.648682 , pass medium(0), pass tight(0)
isolation is: 0.890567 AEff, rhoPrime, iso_nh, iso_em (iso_n), iso_ch, iso:  0.067      15.4463 0.742566        2.43182 ( 2.13949 ) 1.59347     0.890567

Run: 1 , lumi: 26501 , event: 10591623 has electron with pt: 12.461290 , eta: -2.442097, phi: 2.389979 , pass medium(0), pass tight(0)
isolation is: 0.176879 AEff, rhoPrime, iso_nh, iso_em (iso_n), iso_ch, iso:  0.138      16.0234 0       4.41537 ( 2.20414 ) 0   0.176879

IK says:

Sometimes our charged isolation matches, but sometimes it doesn't,
each time you have more charged energy than I.From your events
below, skipping everything below 9 TeV which I don't have in my ntuples:

10591606 - pt = 23.368,   iso_ch: LS 12.9885,     IK 0                   mismatch
10591610 - pt = 29.197,   iso_ch: LS 0.985613,   IK 0.985626      ok
10591623 - pt = 12.461,   iso_ch: LS 0                 IK 0                   ok

are you computing isolation
with PFnoPU candidates?
this is same what he asked last year. at that time after many checks he said "The MIT code does not do literal PFnoPU, but some kind of emulation because of the nature of their framework. It is supposed to be quite close, however. So I will think about it a bit."

I would send him numbers again...need to find time.

Oct 3:

I installed Root 5.34 on my it is interrupting my working codes frown ... will see tomorrow why I am not able to run ExRootAnalysis now frown YM added this to lugg meeting agenda:

Ilya: will move to effective area as mine, rho calculations need to be switched to latest recipe

YM: see/read more to make sure we dnt need to add anything to ggntpls. I met Ntran for discussing if we need something else to add to branches. He is OK with exiting stuff, but when I asked for jet cleaning etc., then he agreed that we should add jet-cleaning module before we do a match of ca8 and ca8pruned jets. So, I added 6 branches saving variables for applying PFJet ID. there are six new vectors added to existing setup ... tested ... ok

ToDo: aspen for car key, Ersh doctr

Oct 2

Got sometime for MG5: I have a sample with SM wgg process and a QGC one with all 5 dim6 operator coffi. to be at 10:

There are so many issues in IK isolation stuff...he is not using sam erho, Eff area numbers as mine...we are so diffrent in calculations for this here: For me the efefctive area numbers are being accessed from EGamma/EGammaAnalysisTools/interface/ElectronEffectiveArea.h . Isolation calculation is done using same formula as you discussed above:

iso = (iso_n + iso_ch) / pt;

where iso_n = std::max(iso_nh + iso_em - rhoPrime * AEff, 0.0);
event: 2128449 has electron with pt: 31.582241 pass both tight n medium, electron with pt: 19.625193 fail medium bcz isolation is 0.197695 n fail tight because os isolation as well as dphi is -0.026695 ( electron is in endcap with eta -1.73809). AEff, rhoPrime, iso_nh, iso_em (iso_n), iso_ch, iso are: 0.067 , 10.3377, 0 , 0.717439 ( 0.0248149 ), 3.85499, 0.197695

and here is what Ilya is getting:

For electron with pt=19.625 and eta = -1.7..
  chIso = 0
  neuHadIso = 0.717438
  gamIso = 0
  rho = 10.951123
  Aeff = 0.07

  iso = (0 + Max( 0.717438 - 10.951123 * 0.07, 0) ) / 19.625 = 0.0 (b/c Max = 0)

Oct 1

YM: unhappy - why thr is no email/report on ggNtuples.

I have problem in accessing nsubjets kinematics only because nsubjet counter branch was not filled properly. Following two variables should give same result (but they don't):

For jth jet:
int nSubJets_one = CA8prunedJet_nSubJets->at(j);
int nSubJets_two = CA8prunedJet_SubjetPt->at(j).size();
So, going through the ntuple maker code, I have tried to change the relevant lines as follows:
//    CA8prunedJet_nSubJets_.push_back( !CA8matched ? 0 : ijetCHSmatch->numberOfDaughters() );
    std::vector<TLorentzVector> subJets;
    int nsubjets=0;
    if( CA8matched ){
      for(unsigned int idaughter=0; idaughter!=ijetCHSmatch->numberOfDaughters(); idaughter++){
        if( ijetCHSmatch->daughter(idaughter) != 0 && ijetCHSmatch->daughter(idaughter)->pt()>0.)
          subJets.push_back( TLorentzVector(ijetCHSmatch->daughter(idaughter)->px(),

Here is what I get with Ming's file:

IK: want to know more about the isolation stuff: will do later.

Need to edit py files for running on data: git clone at lpc, set mc matching bools in =PhysicsTools/PatAlgos/python/producersLayer1/ to false. exclude mc-matching modules from final jet sequence. sent email to CM - edit n commit the changes to git

I went to ntran: he said: they used t2/t1 < 0.5 for H-ww analysis along with a mass cut of 65-105 and for WW analysis they loosen that t2/t1 cut to be 0.55 now, he said this is what we should optimise for our analysis...depending on our choice of high signal purity or high efficiency what is major background for us: dijet and may be we need to concentrate on signal purity.

see this:

If I plot pruned jet mass with in 60-120 mass, ca8prunedJetMass_window2.png , we have 2206 events from 10K ... so 2.2% effi. only.

Sep 30

I am done with my EGamma update:

Need to conc. on CM jet-merging thing. Pick the new sample "job_estar_eejj_m1000.root".

YM helps me with mac - - - done with Xcode -gcc - root installing smile

There is something bad with nSubjet branch...why my loop over ca8 jets when count for nsubjets give me 2 always? Need something in ntuple maker...will see if nsubjet filled properly.

Sep 27, 2013

startd with FCAL meeting: sasha need full dose map in cfc detector in txt format ( sent : ) - dose is in GeV/g - his slides in meeting today include dose map from my simulations ( ), why cfc calculations have that much drop in effi. as compared to EE of original detector- i need to do simulations without PE layer bcz that could be a reason to produce secondaries. So submiited jobs for simulations again without PE layer at lxplus in conf3 directory. I can do this with PE+B-10 layer also. I submitted the jobs for that also ( in conf3withPEB10 layer).

YM discussed ongoing/current issues.

IK: he got new ntuples with proper OfflinePV access and would produce results with same binning as mine for SFs ( use of gsfEle Pt). I am going to start making templates etc. now. but before that IK email says: Lvdp dnt hav 839961 , 7943104, 8927048, 23019670, 28033364, 28065427 (), 45466058, 68422360, 70174878

event: 28065077 has electron with pt: 102.044937 which pass both tight n medium, electron with pt: 18.125261 pass both tight n medium, electron with pt: 6.401584 fail both.

event: 28065427 has electron with pt: 19.145262 which pass both tight n medium, electron with pt: 142.080536 pass medium but fail tight with isolation 0.102131 ( for isolation), electron with pt: 2.295287 fail both.

event: 28033364 has electron with pt: 3.069866 fail both, electron with pt: 35.522766 pass medium but fail tight with isolation 0.102252 ( for isolation), has electron with pt: 18.750225 ( <25 GeV tag requirement)

event: 54509262 has electron with pt: 17.794640 which pass both tight n medium ( <25 GeV tag requirement), electron with pt: 84.117928, d0 -0.0216473 so fails medium and tight both ( for d0 cut < 0.02 )

event: 23019670 has electron with pt: 165.531754 pass both medium n tight, electron with pt: 17.555874 , d0 -0.0448825 so fails medium and tight both ( for d0 cut < 0.02 ), electron with pt: 4.962339 ( very low energy)

event: 68422360 has electron with pt: 72.834244 pass both medium n tight, electron with pt: 14.999915 , sigmaEtaEta 0.0327604 so fail both medium n tight

There 71 events from IK list which I fail, but he pass (he has some loose selection?): 5 of them are: 1251813, 1770393, 1870391, 2128449, 3518872.

event: 2128449 has electron with pt: 31.582241 pass both tight n medium, electron with pt: 19.625193 fail medium bcz isolation is 0.197695 n fail tight because os isolation as well as dphi is -0.026695 ( electron is in endcap with eta -1.73809)

event: 3518872 has electron with pt: 65.616135 pass tight n medium, electron with pt: 16.569344 fail medium n tight because isolation is 0.104465 ( eta: -1.88368 ),

event: 1251813 has electron with pt: 39.532906 pass tight n medium, electron with pt: 19.145323 fail medium n tight as isolation is 0.101165 ( eta -1.56843 ) plus dEta is -0.00530744

event: 1770393 has electron with pt: 36.460499 pass tight n medium, electron with pt: 17.696949 fail isolation 0.104953 with eta 1.82889.

it seems like he has something bad with isolation requirements this time.

Lecture "The Energy Dependence of the Underlying Event in Hadronic Collisions: Tevatron to the LHC " ( Rick Field, Florida, Wilson Hall, room 1 West. )

Sep 26, 2013

Ershleen's b'day

YM: start working on DQM electrons

LPC Workshop on Exotic Top Partners : LPC Physics Forum : A. Martin - Phenomenology of Vector-like quarks, D. Ferencek - Top/Higgs Tagging

YM need - wgg plots from MG5 - dijet mass plot for ggNtuples (file sent for estar-eejj by CM)

Nural, Sasha need Fluka maps : sent email to them ( )

Sep 25, 2013

I did a commit with new cfg file to Ming's git. Ming will produce ntuples for Z->jj.

Interesting stuff particularly FLUKA one:

since I have lhe merging tool, so submitted jobs in parallel.

Sep 24, 2013

work for CM: follow this - account created - goto -

cmsrel CMSSW_5_3_12
cd CMSSW_5_3_12/src/
git config --global <First Name> <Last Name>
git config --global <Your-Email-Address>
git config --global user.github <Your-Just-Created-GitHub-Account>
git cms-addpkg PhysicsTools/PatAlgos
git cms-merge-topic -u vadler:53X-tagset133511
git cms-addpkg PhysicsTools/PatUtils
git cms-merge-topic -u TaiSakuma:53X-met-130910-01 
git cms-merge-topic -u vadler:53X-tagset133511-newBTagging
git cms-merge-topic -u vadler:53X-tagset133511-newEGIsolation
git cms-merge-topic -u cms-tau-pog:CMSSW_5_3_X 
git clone
git clone
git clone -b V01-00
git clone
cvs co -r V09-00-01 RecoEgamma/EgammaTools 
cvs co -r V00-00-09 EgammaAnalysis/ElectronTools 
cd EgammaAnalysis/ElectronTools/data/
cat download.url | xargs wget 
cd ../../../
cvs co -r V00-02-03 -d TSWilliams/BstdZeeTools  UserCode/TSWilliams/BstdZee/BstdZeeTools
cvs co -r V00-09-03 -d SHarper/HEEPAnalyzer UserCode/SHarper/HEEPAnalyzer
cvs co -r V00-11-17  DPGAnalysis/SiStripTools 
cvs co -r V00-02-05 -d CMGTools/External UserCode/CMG/CMGTools/External

cvs co -d JetSubstructure UserCode/ntran/JetSubstructure
cvs co -r V00-06-cmshats2013-08 RecoJets/JetAnalyzers
cvs co -r v1-2-5 -d QuarkGluonTagger/EightTeV UserCode/tomc/QuarkGluonTagger/EightTeV
cvs co -r HEAD -d ExoDiBosonResonances/PATtupleProduction UserCode/ExoDiBosonResonances/PATtupleProduction

for QGtagger twiki says we can do:
git clone; cd QuarkGluonTagger ; git checkout v1-2-5 ;
( )

scramv1 b -j 100
for cmslpc :
cmscvsroot CMSSW
cvs login
 #at the prompt enter the password 98passwd 

Note: file truncated error - you run into disk quota problem

CM need a python file for subjets - done.

For lhe file merging use:

Jet substructure studies in ATLAS by Deepak :

Sep 23, 2013

MG5 jobs were submitted on condor batch system of fnal. Remember : change beam energy, random number -think, nevents.

I submitted jobs for SM as well as EWdim6 models for wgg process...used YM files for Dim6 and MG-tutorial for SM.

Now how to merge all .lhe files?

Sep 20, 2013

YM meeting: slides shown for fcal, we tried to make sense of trends smile

FCAL meeting: , slide5 is interesting, normalize results to say 1/fb and give number to nural to add to his roadmap document: , Alex. Led. asks for Dose maps which I have and would show next week. Alex. mention during his presentation that he may need maps from me - from fluka simulations wink

Sep 19, 2013

Wgg meeting- here are some useful links: and

spent full day on making macros for reading fluka outputs, converting binary fluka output to text or root format. Will show the results in fcal meeting:

Sep 18, 2013

Met T. Bolton smile

Group discussion - attend Wgg meeting tomorrow - Misha testing ntuples, work there

IK would look into MG files for those 2 events where we differ in d0 evaluation. I send him those two files.

MM would edit the ggNtpls today. Ntran came to office to discuss my task. He corrected me: instead of using pruned collection of jets for filling the N-subjettiness as well as other jet kinematics related variables, do it from normal CA8 collection. He further said that although we can do it from both type of collections but CMS prescribe to use following: Provided we have two collections "CA8" and "CA8pruned": - save eta, pt-corrected, tau1, tau2, tau3, area etc. from "CA8" - and jet-mass from "CA8pruned".

Concentrate on Fluka: Collect all usrbin data to plot them. Convert to root format. Here is a neutron fluence plot from new simulations and tools:

neutron f histo 0.1fb noGeo.png

Sep 17, 2013

Met Sadia smile

Yesterday, I moved back to wisc systems for using their condor systems - asked for 5GB user area - jobs done for TTH but failed for QCD because some fatal error frown For a quick update to send to YM, I am submiiting again those QCD jobs but with 1500 events per job wink

Had group meeting - YM need instructions for adding jet-merging stuff - Thurday 9AM-fnal attend Wgg meeting - start using git ( ) - show something in fcal meeting on friday.

I plotted all the jet variables for DY. TTH, QCD etc. Here are the main interesting plots ( but without any mass selection):

h prunedJet tau21.png
h prunedJet tau32.png

Sep 16, 2013

Group meeting: there would be FCAL beam tests, subscribe to ecal-hcal upgrade group, will meet in mid-week for sharing ggNtuples updates.

I added tauN (N=1,2,3) varibles to ggNtpls ( ref. , ). A nice snap-shot for jet-merging:


Spent 3-4 hours on learning crab job submission on fnal systems. All the tries were either terminated or with some job exit status. In the end, emailed cms-t1 which says "resilient" area is full, so my output not stored there. users are requested to remove junk files from their areas frown

Sep 13, 2013

Aimed for MG5 today, but YM emailed for jet-merging idenification and meeting with ntran - if he find time to reply then meet him. Purpose of this meeting would be to learn using jet-merging tools, in other words, YM want to add some variables like #wideJets, #subjets etc. in ggNtpls so that we can use them for analysis lateron. He asked to make sure we have all appropriate variables in trees and all relevant objects are accessed. I did something similar in June first week: Summary of brief meeting with NT: seems happy with present implementation, add N-subjettiness variables tau1, tau2, tau3 and pruned jet mass to ntuples, prescription is to cut on these two variables. Keep ca8 and ca8pruned jet kinematic variables in tree and make sure to see dR matching between these two new jet-collections. Need to read , shift working area from wisc to fnal, everything working fine there.

I accidently removed my fluka fluence results...need to resubmit jobs frown ... done this at lxplus (directory is Configuration3). Did some simplifications to cms geometry files and added emf thresholds on new regions that i defined.

Started using Desktop at office, laptop is dying.

IK asked same information as was asked in 2011 wink - what is the name of the PV collection you are using? Ans: primary vertex input tag is: "offlinePrimaryVertices" - which PV out of that collection are you using as a reference to compute d0? Ans: We pick 0th position vertex for computing d0 - what exactly function are you calling on a track to find its d0? Ans: d0vtx = ele->gsfTrack()->dxy(vtx->position());

Sep 12, 2013

Sent email to for help regarding public_html. Working now: see my MG5 work from yesterday:

Reading Fluka-manual: USRYIELD is another good option for scoring fluences as function of various energy like quantities, but again all these options need volume/area for region of interest. Is there any tool for these calculations? ... sent email to Moritz and Anne...need some expert in FLUKA at LPC frown

Viktor chat: want to write note, but from whatever ws already done in past and within this month ... seems tough for me...but I also want to conclude over results in an AN at least.

Moritz go for aim for WindowsOS.

Need to go for ES&H warning again frown - Job Hazard Analysis, Traffic Safety Awareness: done most easily, rest of them has taken my 3-4 hrs - but finally I am done with that.

Email Ilya for SFs. Got reply: busy, noupdate

Couldn't find WindosOS today. But here is what I got from my last jobs which were submitted for USRBDX. This plot gives neutron and photon spectra at various surfaces: one is W-86 to air surface and another PE to W (there we put PE layer of 20cm thickness before W-86). See slide-3 of my update in FCAL meeting - Aug 2


Sep 11, 2013

FLUKA jobs for USRTRACK and USRBDX still running, meanwhile I can have a look at MG5 instructions from Yurii. Ilya (on skype) still has not checked his d0 values, so I will wait for him.

TODO: fermilab ES&H tests ... they have sent thousands warnings frown

MG5 - copied all the stuff to my area - vi this is replacing Nevents n EBeam in original run-card + replacing dim6 operators in param-cards with your choice

  • run-card stores beam and run parameters and also cuts, param-card has model parameters
  • prepare run_waa*.dat which import EWdim6 model , run mg5 using makeJobs.csh and submit them on condor using submitJobs.csh
  • I am trying to test it locally - so do python lv 1000 4000 10 10 10 10 10 , this gives waa_lv directory and run_waa_lv.dat file - next step is do: $ ./bin/mg5 < run_waa_lv.dat
  • running - so many things - but in end it gives a index.html, which I am sending to wisc for reading now ( I need to see how and where to make public_html at cmslpc computers): , 1152 diagrame wink - this has complete process production information - seems very useful - read more...but I guess such weblink would be useful only if I have it on my cmslpc account ( TODO: ask cms-t1 or for help)
  • cd waa_lv/Events/run_01/ , gunzip events.lhe.gz
  • ExRootAnalysis - convert lhe to root using:
./ExRootLHEFConverter events.lhe mg5First.root
** Calculating number of events to process. Please wait...
** Input file contains 3517 events
[#########################] (100.00%) : 3517/3517 entries processed
** Exiting...
mg5First.root has LHEF tree and here is a quick look at particle-Pt variable from tree:

Sep 10, 2013

FLUKA: Yesterday I submitted some jobs for a boundary crossing fluence, I defined the score card like this:
USRBDX 99. NEUTRON -36. R9991L R373 neutWtoAir
USRBDX 1E-9 10. &
  • R9991L is a region in cms geometry file which I defined for W-HF at z=322.3 - 442.3cm and R373 is a region which has air now ( but before that in original geometry it was having layers of HCAL). In this card, I am setting detector volume to be default ( 1cm^2) for now ( will read manual for this - doubtful)
  • jobs are still running.
I also added USRTRACK score-cards yesterday, in separate jobs which are running fast being on 2nd queues (may be)- this help to calculate fluence as a function of energy - it scores track-length in given volume ( which I am again setting to default - 1cm^3 )
USRTRACK -1. NEUTRON -32. R9991L 10.NFTrack
Defined same for photon and for some of the events (~5K), here is what I am plotting:
NeutronPhoton logEnInt 10bins.png
TODO - read manual or ask someone to make sure if I need to put correct volume here ?)

TnP SFs: Merged the madgraph MC trees which have tau-filter variable as well as mcTrue matching done. Copy this to hdfs for template making. Done with template making. Here are my MC templates: I asked Ilya for his templates and shape comparison. He sent his Sc-Et templates in root files but those templates which are based on Ele-Pt have pass/fail events in root format - two trees storing basic vaiables. If I need to plot all these templates, play with root.

There was a meeting "new Fermilab director: Nigel Lockyer" - discussed money + future of big projects or say our community.

Edit | Attach | Watch | Print version | History: r63 < r62 < r61 < r60 < r59 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r63 - 2013-11-26 - LovedeepKaurSaini
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2021 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback