Comparison fast off-line (dev, 12/28/2019) and TFG19m (new alignment and calibration)

 The express TPC alignment and dE/dx calibration based on Cosmic and 11p5GeV collision data for Run XX has been finished 2019-12-24 and frozen as release TFG19m.
This alignment and calibration have not been put in DB awaiting comparison between Fast Off-line and TFG19m.
In TFG19m has been processed ~55 M event for  11p5GeV and ~1 M 5p75GeV_fixedTarget.
The location of this production on RCF is the following:
   du -sh  /gpfs01/star/pwg_tasks/tfg02/2020/TFG19m/RF/*GeV*

  • 3.3T    /gpfs01/star/pwg_tasks/tfg02/2020/TFG19m/RF/11p5GeV.B
  • 29G     /gpfs01/star/pwg_tasks/tfg02/2020/TFG19m/RF/5p75GeV_fixedTarget.B

Using a sample ~10 K event obtained in dev and TFG19m we have compared momentum reconstruction using Track-By-Track comparison
and found difference ~1%/1GeV in pT measurement.
See https://www.star.bnl.gov/~fisyak/star/dev_TFG19m_comparison/Inv_pT_diff_perc_vs_pT_for_Global.png  and https://www.star.bnl.gov/~fisyak/star/dev_TFG19m_comparison/Inv_pT_diff_perc_vs_pT_for_Primary.png
The difference is due to the alignment parameters only. The difference in codes (dev or TFG19m) is not visible.
We have  tries to understand what calibration is better. For that we used pull between measured and fitted parameters for V0 reconstruction.
They are summarized on  https://www.star.bnl.gov/~fisyak/star/dev_TFG19m_comparison/index.php

where we compare dev (~8.9 M events) with TFG19m (~42 M event) (https://www.star.bnl.gov/~fisyak/star/dev_TFG19m_comparison/PrimaryVertexQAz.png)
The conclusion

  1. y of primary vertex is shifted by ~0.3 mm (https://www.star.bnl.gov/~fisyak/star/dev_TFG19m_comparison/PrimaryVertexQAy.png)
  2. Background for Ks in dev higher than in TFG19m : https://www.star.bnl.gov/~fisyak/star/dev_TFG19m_comparison/KsM.png
  3. Pull in Ks mass in dev is worse than in TFG19m : https://www.star.bnl.gov/~fisyak/star/dev_TFG19m_comparison/KsPullM.png
  4. ...