The result is that when you iterate over contiguous memory, you can access it about as fast as the CPU can operate, because you will be streaming chunks of data

The result is that when you iterate over contiguous memory, you can access it about as fast as the CPU can operate, because you will be streaming chunks of data into the L1 cache. This is not true at all. You'll be able to access it about as fast as the memory can operate - it'll still be much faster than randomly chasing pointers all over the place - but the maximum processing speed of the CPU is an order of magnitude greater than that again.