Mokka vs. LCDG4 Comparison

1
Mokka vs. LCDG4 Comparison

• Guilherme Lima
• DHCal Meeting Oct. 1st, 2003

2
Comparing apples to apples

• Geant version (4.5.1 in nicadd, 4.5.2 not
  installed)
• Geometry (SDJan03, projective)
• Physics list identical data files (almost)
• Range cuts (0.7mm)
• No energy thresholds
• Geant4 external data files not used (dont
  matter)
• Text output implemented into projective LCDG4
• Angular distributions (input from file, bad
  random number generators) theta 90deg , flat
  in phi
• Materials (some differences, maybe ok for cals?)
• No bugs!!!

3
Reconstructed energy 50 GeV e
4
Reconstructed energy 50 GeV pi
5
Ecal Energies per layer
6
Hcal energies per layer
Problem?
7
Ecal energies in absorbers
8
Hcal energies in absorbers
9
Ecal energies in absorber CELLS
10
Hcal energies in absorber CELLS
11
Ecal energies per cell
Excess!
12
Hcal energies per cell
Different slopes
13
Ecal number of hits
14
Hcal number of hits
Thresholds At 700 keV
15
Number of hits per layer
16
Cross checking EcalHcal, abscell
17
Conclusions

• Some low-level discrepancies in both ECal and
  HCal (materials?!)
• Discrepancies to Gismo were at much much higher
  level
• Should we go for Geant3 cross checks?
• Need better sampling fractions (Vishnu?)
• Need to understand peaks at ECal absorbers

18
Next steps

• Still some checks needed for reliable
  comparisons
• Better random number generator for kinematics
• Energy 50 GeV, not momentum
• Make sure materials used are the same
• Input from file, for processing exactly same
  events
• Document plots in terms of CVS tags
• Investigate origin of discrepancies