A lot of people look at sorting algorithms with the mental model of a 1970s era PDP-8. They analyze scalar instructions as if they are being executed sequentially.

Once you admit superscalar execution (multiple loadstore pipes), out of order execution and caches into the picture ( as well as the possibility of almost sorted arrays as input) the picture is never as simple.

That said, I’m curious whether insertion sort is categorically better than bubble sort. I thought they might have been equivalent for the almost sorted case. I’ve heard both sides argued on this thread.