Virtual Memory

Question 1

what is the motivation behind virtual memory?

Answer 1

• logical and physical addresses are completely separated
• processes address a virtual address space, which is larger than physical memory
• so programmers don’t have to be concerned with physical memory size
• more programs can run at the same time because they’re not fighting for space
• less I/O required to swap programs in and out of memory, increasing performance
• the entire program doesn’t have to be in memory
• allows shared memory

Question 2

what is demand paging (basic concept)?

Answer 2

the entire program may not be needed when running. when using demand paging only the required pages are loaded into memory, as they are needed

Question 3

what is a lazy swapper?

Answer 3

only swaps pages when they are needed

Question 4

what is the purpose of the valid/invalid bit?

Answer 4

each entry in the page table has a flag bit that indicates if the page is in memory
if the process tries to access an invalid page, we trap to the OS in a page fault, which loads the page from disk and restarts

Question 5

what the 5 steps in a page fault?

Answer 5

1. check if address is legal
2. find a free frame
3. schedule disk read operation
4. set the frame number and valid bit in the page table
5. restart the instruction

Question 6

what is pure demand paging? problems?

Answer 6

we dont have any pages in memory when we start execution, the first instruction triggers a page fault. pages are only loaded via page fault. restarting instructions can be expensive

Question 7

how do we calculate the EAT in page faults?

Answer 7

p = probability of page fault
ma = memory access time
EAT = (1 - p) x ma + p x page fault time

Question 8

what is the bottleneck in page faults?

Answer 8

disk reading because its mechanical, there’s competition

page faults REALLY slow things down

Question 9

what is and wat are three strategies for handling over-allocation?

Answer 9

• terminate the process
• swap out a process entirely
• page replacement!

Question 10

what is the page fault routine with replacement?

Answer 10

• find location of page on disk
• look for a free page frame
• if free, carry on
• if not, use a page replacement algorithm to select a victim
• copy victim to disk and mark as invalid
• read the page into the frame

Question 11

downsides of page replacement? how to mitigate?

Answer 11

• 2 page transferes are neeeded, which increases our page fault time
• unmodified pages dont have to be written back, so we can use a dirty bit indicating wheter the page has been modified since it was faulted