*** Discussion title: Tracker Alignment

Dear all,

since I am not available the next two weeks, I want to give an update of 
the current preparation for the MC alignment scenario for 2011.

Finally, I was able to run the alignment after reducing the number of 
cosmics tracks in order to decrease the number of non-zero off-diagonal 
elements in the alignment procedure. In total three alignments were 
performed which means that the alignment was iterated twice. Therefore, 
compared to the status I presented in the last tracker alignment meeting 
(https://indico.cern.ch/getFile.py/access?contribId=8&sessionId=1&resId=0&materialId=slides&confId=148985) 
the weights for the individual data sets had to be adjusted according to

2872412 cosmics 1.9311504944
3361479 WtoMuNu 6.9242266535
6896912 QCD_Pt-20_MuEnrichedPt-10 0.2260290298
6108415 MinBias 0.4815410822
554313 ZtoMuMu 1.0                              

where the first number corresponds to the number of tracks or z->mumu 
candidates, respectively, the second string represents the data set 
name, and the last number the applied weight in the alignment procedure.

The resulting geometries can be found here

first alignment:
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/MP/MPproduction/mp0860/jobm_it0/alignments_MP.db

iterated alignment:
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/MP/MPproduction/mp0860/jobm_it1/alignments_MP.db

2. iteration:
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/MP/MPproduction/mp0860/jobm_it2/alignments_MP.db

Geometry comparisons and offline validations were also performed. 
Unfortunately, due to a not understood technical problem the latter was 
only determined for the first alignment, but it should be easily 
possible to rerun everything because I will provide in the following all 
necessary config files (the default ini-file that includes the sensor 
deformations has to be used). In MC the offline validation was fed with 
the WtoMuNu sample whereas for the data the TkAlMuonIsolated stream was 
used which means that the track sample is not completely consistent 
betweem data and MC. In addition, however, for data only runs from IOV 
period G were considered.

first alignment:
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/results/jbehr/mcalignment2011_withdeformations

I hope the names of the geometries are clear.

In the subfolder ExtendedOfflineValidation_Images/ DMR plots can be found.

The geometry comparisons for the other two alignments are stored in 
these folders:

/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/results/jbehr/mcalignment2011_it1

/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/results/jbehr/mcalignment2011_it2

The used configs are (In this files you can find comments how to 
resubmit them):

/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/alignmentObjects/jbehr/mc_alignment.ini
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/alignmentObjects/jbehr/mc_alignment_it1.ini
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/alignmentObjects/jbehr/mc_alignment_it2.ini

Finally, the fully parsed config output of the jobs:
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/alignmentObjects/jbehr/mcalignment2011_withdeformations/
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/alignmentObjects/jbehr/mcalignment2011_it1/
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/alignmentObjects/jbehr/mcalignment2011_it2/

In order to create the DMR plots I have used this macro:

/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/alignmentObjects/jbehr/mcalignment2011_withdeformations/TkAlExtendedOfflineValidation.C

In that file you can also find the path to the other plotted root files 
(data, MC ideal, MC 2010Relastic). In order to increase the statistics 
for data it might be useful to merge all the IOVs. Therefore, I have 
prepared the offline validation root files also for the other IOVs:

/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/results/jbehr/paper_data_*

The wildcard * represents a letter from B to I. I hope that these jobs 
were configured such that all module level residuals are included.

The impact of the new MC alignment on the dimuon mass was also studied:

/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/results/jbehr/mcalignment2011_it2/mz_eta.eps
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/results/jbehr/mcalignment2011_it2/mz_phi.eps
/afs/cern.ch/cms/CAF/CMSALCA/ALCA_TRACKERALIGN/data/commonValidation/results/jbehr/mcalignment2011_it2/mz_phi_iterations.eps

In the latter file the results for the three different iterations are 
compared. One can see that small improvements are visible. However, the 
alignment seems to have converged. Compared to data the size of the 
amplitude is almost twice as large and the dependence on eta is stronger.

I hope thats all you need.

Best regards,
Joerg

-------------------------------------------------------------
Visit this CMS message (to reply or unsubscribe) at: 
https://hypernews.cern.ch/HyperNews/CMS/get/tk-alignment/1103.html