Difference between revisions of "HDGeant4 Meeting, January 28, 2020"

From GlueXWiki
Jump to: navigation, search
(Adding minutes)
m (Systematic studies of Geant3/Geant4 differences)
 
Line 38: Line 38:
 
=== Systematic studies of Geant3/Geant4 differences ===
 
=== Systematic studies of Geant3/Geant4 differences ===
  
Alex presented Igal's work on Monte Carlo studies of &gamma;p&rarr;&omega;p and &omega;&rarr;&pi;<sup>+</sup>&pi;<sup>&minus;</sup>&pi;<sup>0</sup>;. He compared simulations with:
+
Alex presented Igal's work on Monte Carlo studies of &gamma;p&rarr;&omega;p and &omega;&rarr;&pi;<sup>+</sup>&pi;<sup>&minus;</sup>&pi;<sup>0</sup>. He compared simulations with:
 
* G3 vs. G4
 
* G3 vs. G4
 
* no accidental background events, background from Richard's MC beam background generator, and random trigger events mixed in
 
* no accidental background events, background from Richard's MC beam background generator, and random trigger events mixed in

Latest revision as of 11:51, 12 February 2020

HDGeant4 Meeting
Tuesday, January 28, 2020
3:30 pm EST
JLab: CEBAF Center, A110
BlueJeans: 968 592 007

Agenda

  1. Review of minutes from the last Meeting (all)
  2. Systematic studies of Geant3/Geant4 differences (Igal)
  3. Issues on GitHub
  4. Pull Requests on GitHub
  5. Action Item Review

Minutes

Present:

  • CMU: Naomi Jarvis
  • FSU: Sean Dobbs
  • JLab: Alex Austregesilo, Colin Gleason, Mark Ito (chair), Simon Taylor, Beni Zihlmann
  • UConn: Richard Jones, James McIntyre
  • W&M: Justin Stevens

There is a recording of this meeting on the BlueJeans site. Use your JLab credentials to get access.

Review of minutes from the last Meeting

We went over the minutes from January 14.

Calorimeter timing mismatch between g3 and g4

This is Issue #93.

Colin has extended his work on this issue, widening the timing cuts as Alex did for his ρ analysis. He sees indeed that G3 and G4 both have significant tails in BCAL timing for πs, with the tail bigger for G3.

When looking at acceptance for the full reaction (pγ→ηπΔ++) as function of Gottfried-Jackson angle the differing shapes remain, except the G4 efficiency drops in level from about 4% to 2.5% while retaining its overall shape. Alex saw a similar effect in the ρs until he fixed the bug whereby the choice of RF bunch was defeated by a lack of agreement among multiple tracks inside the wide timing window. Colin will go back and repeat his study with the current master branch of halld_recon.

Systematic studies of Geant3/Geant4 differences

Alex presented Igal's work on Monte Carlo studies of γp→ωp and ω→π+ππ0. He compared simulations with:

  • G3 vs. G4
  • no accidental background events, background from Richard's MC beam background generator, and random trigger events mixed in
  • run 30496 vs. run 30730

for a total of twelve different sets of conditions. He showed a variety of different comparisons among these conditions including single particle distributions, invariant mass distributions and occupancy plots for the FCAL. Please see his slides for the rich variety of comparisons available. From Igal's last slide, here are his conclusions:

  • Low-level systematic comparisons between different Geant versions and runs, and with and without background combined: Geant3 and Geant4 show clear difference without any backgrounds included
    • More photons/hits are produced in Geant4 (observed mostly in FCAL due possibly to ω physics/decay)
  • All the other differences listed below are happening outside of Geant Without backgrounds (but also with backgrounds)
    • Calibration tables although identical are applied differently to Geant3 and Geant4 simulations possibly because they are not applied at the same variable(s)?
  • Combination of background/noise to simulation produces different outcome for Geant3 and Geant4 simulations
  • Combination of background/noise to Geant3 simulations are consistent and evenly distributed => seems correct
  • Combination of background/noise to Geant4 simulations are inconsistent and not evenly distributed
    • Background/noise mostly located at the edge and
    • Creates spikes
    • => seems buggy
  • As the same code is used to combine the background/noise to both simulation version, it probably means it is not using the same variable(s)?

Collaboration Meeting

We need to make some sort of report out of this group on work since the last collaboration meeting. Mark and Sean will get together and discuss options. Stay tuned.