⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 48.39%. Comparing base (13ca1b9) to head (f974927).

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

@@            Coverage Diff             @@
##           master     #152      +/-   ##
==========================================
- Coverage   48.53%   48.39%   -0.14%     
==========================================
  Files          17       17              
  Lines        1498     1498              
==========================================
- Hits          727      725       -2     
- Misses        771      773       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

If I'm reading this right, the question the graphs try to answer is "How much faster is using a BVH over bruteforcing for the same parameters". I'm not sure I'm parsing this right but I don't think the speedup would be maxing out at a factor of two since we should be looking at exponential gains over a bruteforce implementation.

Yeah, the speedup has no upper bound. It's currently capped at 2 because it made the color scheme simple, but that could be changed.

Capping at 2, in a clearer way, could still be useful since it shows when a BVH is faster (albeit not how much faster).

I think the current way is a bit misleading. In general, I'm not sure how readable this three-dimensional graph can be made considering colors doesn't map so well to logarithmic scale (or at least it's not intuitive). How about a set of two-dimensional graphs that show a logarithmic speedup?

Also as a general thing, I might be cool to use plotters here 🦀. Not sure how practical it would be, though. :)