I see how with this approach we are kind of interweaving various parts of the array per computation instead of getting a contiguous window to perform an operation on at each time step. Is the approach of getting a contiguous window of the array and loading that always more beneficial to speedup? Or are there times where that would not be a beneficial thing to do and this slide's approach better.

I think it depends on the data inputs, as seen in programming assignment 1.

What is the benefit of assigning instances to consecutive array elements as opposed to separate elements?

When we use foreach, is the behavior of how program instances assign to loop iterations completely undetermined/random?