I’ve tried a couple of fori approaches, and both have been a little over 10x slower than the standard for loop on my benchmark, which is iterating over and summing the elements of a 4*4 matrix (as defined in this post). I’m very curious to see your implementation.

Great stuff! Regarding your fori implementation that is a lot slower–how much slower is it?

I’ve played around a bit with various fori implementations. My initial (naive) approach was almost 10x slower than [for] itself. I’ve improved to where I’ve now got a version that only takes 1.5x as long (on a range of 1-1000 my [fori] takes 1514 microseconds compared to 1007 microseconds for [for]).

Happy New Year.