[Sirened]

It depends on what prefetchers your CPU has and the actual underlying access pattern your accesses cause. If chasing next leads to a constant stride run through the array, you'll get identical performance to that of an iterative walk since essentially every high performance CPU since like 2010 supports stride prediction based on raw addresses. If .next sends you bouncing all over the area, you'll get complicated behavior since whether or not the data can be prefetched depends on the performance/existence of a pointer prefetcher, which is less common/more error prone. We know Apple's M1 have it due to some researchers using it as a side channel [1] but you'll have to do some digging on whether or not your laptop has one. Would make a nice post here if you do make the benchmarks :)

[1] https://www.prefetchers.info/augury.pdf

[raible]

Thanks for that, it's what I had hoped for (but again, have not yet measured).

It seems to me it's a super-handy way of "modifying" a compiler-allocated array of structs. I'm sticking with it!