Suppose we have a range of memory values that are mapped (but not loaded), and we would like to access them. Is there any kind of optimizations the kernel does that would make the sequential access faster than the random access when page faults occur and the kernel loads them into memory?
E.g. if we have 100 pages worth of memory mapped, sequential access will read page 1, 2, 3, ... while random access will read page 10, 2, 5, ... in a random manner. What could possibly account for the difference in latency (if any)?