Integrate with LArSim via standalone plug-in

[sethrj]

https://github.com/LArSoft/larsim is used by DUNE (#1344 ) and other neutrino physics-oriented detectors:

• ICARUS https://icarus.fnal.gov/
• SBND https://sbn-nd.fnal.gov/
• $\mu$ BooNE https://microboone.fnal.gov/

This represents "phase 1" of our integration plan: rather than running simultaneously with Geant4, we take the pre-processed hits (with the number of photons already calculated) and determine the detector responses in a standalone simultion.

Execution plan

See google doc for more details.

• LArSoft cmake and tool interface Add LArSoft plugin skeleton #2154
• Add optical propagation interface to larsoft Add new optical propagation interface that will allow optical transport on GPU nuRiceLab/larsim#1
• Refactor LarSoft PDFastSimPAR_module to call optical propagation module with plugin tool (@stognini )
• Test Celeritas build alongside MRB and calling it as part of a larsoft/dune workflow once the OpticalSimInterface module is complete (@sethrj )
• Update LarStandaloneRunner to build and run an optical Transporter (see Add standalone optical runner #2197)
• Test that we can build the necessary optical physics after a geant4 run manager has been torn down
• Collect and sort photon hits: per detector, sorted by time: use per-detector OBTRHelper class for short term, be smart about it for long term

[sethrj]

Note from our discussion Friday that LArSoft currently uses Geant4 10.6 but is encouraging its users to move to 11.4 .

[sethrj]

Test system setup: #2153

[stognini]

Generating MC samples

With help from Aaron, this is the workflow to generate example Geant4 ROOT files with art energy deposition objects.

• dunesw fcl job files:
  • GENIE: prodgenie_nu_dune10kt_1x2x6.fcl
  • Geant4: standard_g4_dune10kt_1x2x6.fcl
• fcl files are part of the environment once dunesw is set up and thus can be called from anywhere

Generate GENIE events

$ lar -c prodgenie_nu_dune10kt_1x2x6.fcl -n 10 -o genie_output.root

Generate energy depositions from Geant4

$ lar -c standard_g4_dune10kt_1x2x6.fcl -n 10 -s genie_output.root -o g4_output.root

• See standard_g4_dune10kt_1x2x6.fcl

• The PDFastSim producer will likely be replaced by the new OpticalPropagation producer module in the above fcl:

  # Swap the 1x2x6 pdfastsim configuration
  physics.producers.PDFastSim: @local::dunefd_1x2x6_pdfastsim
  

[Rashika-Gupta]

Dune geometry generated using NuRice by is in benchmarks

[sethrj]

@elliottbiondo could you fire some tracks through that geometry?

[elliottbiondo]

Oops, sorry I am just seeing this. I already responded on Slack, but here it is for posterity.

[stognini]

Running PDFastSimPAR module

The default PDFastSimPAR.fcl does not include detector-specific parameters used by ScintTimeLAr and SemiAnalyticalModel tools. The needed parameters are

VISTiming
VUVHits
VISHits

These can be found in the PDFastSim_dune.fcl:

• Values are defined in the DUNE HD 10kt and DUNE HD 10kt 1x2x6 sections.
• Make the data available in larsim (e.g. copy opticalsimparameterisations_dune.fcl to larsim/larsim/PhotonPropagation)
• Update PDFastSimPAR.fcl to point to those parameters:

  #include "opticalsimparameterisations_dune.fcl"
  
  ...
  
  VISTiming: @local::dune_vuv_timing_parameterization
  VUVHits: @local::dune_vuv_RS100cm_hits_parameterization
  VISHits: @local::dune_vis_RS100cm_hits_parameterization
  

• Run the module:
  (See Generating MC samples above if you need g4_output.root.)

$ lar -c PDFastSimPAR.fcl -s g4_output.root -s fastsim_out.root