This article is a great example of the very tangible effects caused by hardware properties that are often disregarded. Very interesting read! Also, I can't hel

This article is a great example of the very tangible effects caused by hardware properties that are often disregarded. Very interesting read! Also, I can't help but note the (probably non-existent) connection to 3-way partitioning for quicksort (which is consistently ~10-15% (iirc) faster than traditional quicksort). It is interesting how we seem to assume that surely factors of two must be the best choice, computers working in binary and all. But sometimes this is true only in theory, but not in practice (binary search), sometimes it's good but not optimal (quicksort), and sometimes it outright is the provably worst choice (dynamic arrays)!