You can't just access memory in that pattern because an out of order processor will just issuse the next N loads in parallel. Instead lmbench sets up a linked

You can't just access memory in that pattern because an out of order processor will just issuse the next N loads in parallel. Instead lmbench sets up a linked list in memory ans walks the list. That way the address for each load is not known until the previous load completes. -Wayne