I feel like if you already know exactly how you're about to access the data, then it would be more efficient to be able to just tell the kernel rather than maki

I feel like if you already know exactly how you're about to access the data, then it would be more efficient to be able to just tell the kernel rather than making it figure it out on its own. At the minimum, the first accesses would perform better since it wouldn't have to waste time learning. This is why instructions like dcbz exist on POWER - to tell it to not bother reading in the next cache line because you're about to write to it.