Question:
Consider a paging system where the page table is stored in the translation look aside buffer (TLB). The hit ratio is 99% meaning the page table entry will be found in TLB 99% of the time. The normal memory access time is t = 1.3 microseconds whereas TLB access is 0.15 microseconds.
a. When calculating the effective paged memory access time, why is the cost of a TLB miss the sum of TLB access time plus normal memory access time?
b. If we consider only swap-in/swap-out time and if swap-in-time = swap-out-time = 12 milliseconds, and, on the average, 85% of the pages are dirty, what is the effective page fault service time?