Rdm

To simplify this further, though: code which accesses memory sequentially tends to have an order of magnitude speed advantage over code which accesses memory randomly. A rule of thumb here, <u>if you're not prepared to spend a lot of time benchmarking</u>, is that aan algorithmic performance advantage which is not at least a factor of 2 is probably a waste of time. (And, in some contexts -- especially where the underlying system was designed to solve a different problem -- factor of 1000 or greater performance improvements might be practical, for reasons independent of cache management.)