Closed cospectrum closed 4 months ago
Thanks for the PR. Does replacing a nested loop with Image::from_fn have any impact on the image filtering benchmarks?
Image::from_fn
iterates in the same order (row-major, cache friendly), so it should be the same. Plus, you don't have to think about the row-major vs col-major every time.
Image::from_fn
iterates in the same order (row-major, cache friendly), so it should be the same. Plus, you don't have to think about the row-major vs col-major every time.
Hmm, not really. A small regression. It seems that the compiler cannot optimize the mutable lambda
Or the iterators are to blame.
A couple of minutes. I'll refactor incrementally
@theotherphil, Now I think my bilateral_filter
bench is better than the master
(both allocs and speed).
Great, thanks!
The only public change:
convex_hull
.