Minutes for LArSoft coordination meeting on Mar 12, 2019

Present | Lynn Garren, Robert Hatcher, Alex Himmel, Kyle Knoepfel, Saba Sehrish, Erica Snider|

Remote | Patrick Green, Diego Garcia-Gamez, Andrzej Szelc, Katherine Lato |

Release and Project status report [Erica Snider]

• v08_12_00 released on march 6
• this week updates to nutools, and no meerge requests so far
• Neext week:
  • migration to root v6.16/00 - to fix an iorule problem
  • will combine with migration to art 3.02
• Genie v3
  • validation testing by MicroBooNE, once done will get sign off from other experiments
• LArTPC data overlay workshop
  • Meeting notes and conslusions are available
• LArSoft workshop 2019
  • 1.5 days, dates TBD
  • Tentative program - LArSoft introduction, Advanced computing techniques, dnd Long-term vision
  • [Lynn] suggested to add a session patch releases

Conclusion:

A semi-analytic way of Simulating light [Andrzej Szelc]

• Introduces analytic model for light propagation
• Optical libraries size are very large (for DUNE and SBND, > 1 GB)
  • problem for cvmfs, so moved to StashCache
• Does not provide timing information (though this is solved in LArSoft)
• Need to run a campaign of grid jobs every time a detector parameter changes
• For DUNE 1x2x6, segmenting the bars into sub-detectors makes it impossible to generate the library due to memory issues. Needed for the TDR
• DUNE needs realistic X-Arapuca supercell geometry for physics studies. SBND libraries also becoming cumbersome
• An alternate method for simulation of light to make things work
  • Improved parameterization of photon arrival times for both VUV and visible light
  • A semi-analytic model for predicting N hits based on position in the detector for direct VUV and reflected light coming off cathode
• Simulation of the VUV arrival times
  • this work provides optimzation and extension of existing feature in LArSoft for the photon arrival time
  • VUV arrival times change as function of distance due to Rayleigh scattering
  • Index of refraction also changing rapidly across emission spectrum
  • This needs to be taken into account
    • Landau + exponential for short distances (< 300 cm)
    • Landau for longer distances (> 300 cm(
  • Visible arrival calculated as distance / velocity
  • Smearing in times approximated as exponential (w cutoff)
• Semi-analytic modelling of light (direct)
  • Calculate solid angle
    • Functions for square and circular detectors
    • No library needed, so avoid memory issues
    • No issue from segmentation of bars into X-Arapuca supercells or windows
    • Can scale to full size of DUNE
  • Need correction due to Rayleigh scattering + reflections
  • Detailed study of border effects on-going
• Corrections to direct light due to scattering
  • Gaisser-Hillas functions fit well
    • Based on ratio of solid angle calculation to full simulation as a function of angle from photo-detector
    • Parameterize as function of angle and distance to photo-detector
• Show agreement with full simulation is better than for libraries
• Reflected light (visible)
  • Created by TPB coated foils on cathode
  • Calc cathode solid angle for VUV
  • Correct for scattering using Gaisser-Hillas curves
  • Detected photons from solid angle to optical detector (from foils)
  • Parameterize corrections as function of angle, distance from cathode
• VU direct light results:
  • there is no bias and performs better the photon librarries
• Performance for DUNE 1x2x6: also better than the optical libraries
• Should not be looking outside the range ( for range -500 < y < 500 cm)
• Most of the changes in larsim
  • bulk of code is in LArG4/OpFastScintillatiton
• Improvement in the PhotonVisibility Serevice
  • use pre-calculated values result in huge efficiency improvement
  • discrepency introdiced by this is negligible ~0.1 ns
• the use of feature is off by default
• fcl configuration provided to enable optical library and time correction and also to enable nhits and time correction
• More improvements to the code: remove duplicated code
• Future improvements:
  • is it worth splitting the photon counter module into two since it does too many things
• based on old LArG4 - for DUNE
• Testing
  • DUNE CI tests
  • ArgoNeuT and LArIAT ok
  • ICARUS and MicroBooNE have some problem
  • SBND failed. Wrong tag. Will re-check
• Almost ready to go (modulo the failing CI tests)
• Thread safety?
  • Not
  • Should be able to make it not a service (Photon visibility)
• Kyle raised some issues related to multi-threading
  • Will address with the move to the new LArG4.

Conclusion

• This feature request was approved.
  • larsim: feature/lightprop_ugr_mcr
  • larana: feature/lightprop_ugr_mcr