Testing proper definition of geometry for raytracing with ELV scanning

[ggalgoczi]

Simon recommended a specific way to understand if our geometry is well defined in terms of performance for raytracing.

https://groups.io/g/opticks/topic/performance_test_opticks_cats/106900329

[ggalgoczi]

"To find problem geometry I have a procedure (ELV scanning) that dynamically changes solids included/excluded within the geometry measuring ray trace render performance for each using CSGOptiX/cxr_scan.sh
As simulation performance is limited by finding intersects fast rendering corresponds to fast simulation.

The render scan results allow you to create a table that orders all the geometry solid inclusions/exclusions by their render times. If the difference between the fastest and slowest is reasonable (<~5x) then you can say you do not have any extremely bad solids. I plan to do this again for JUNO during July. I will write some notes documenting how to do this so others can try to do the same."