For 2/23

Question 1

What does virtual memory allow for?

Answer 1

Parts of a program, only the current instructions and data, need to be in memory for execution. The rest gets a virtual address in secondary memory

Question 2

Address space

Answer 2

The range of memory addresses available to a process

Question 3

virtual address space

Answer 3

The virtual storage assigned to a process

Question 4

Real address

Answer 4

The address of a storage location in main memory

Question 5

resident set

Answer 5

The portion of a process that is actually in main memory

Question 6

What happens if there is a page fault?

Answer 6

The process is blocked, what is needed is fetched while other things are processed, and then it is ready again

Question 7

What is an advantage to only having a resident set?

Answer 7

More processes can be in main memory at a time, and a process can be larger than all of memory

Question 8

Real memory

Answer 8

The same as main memory

Question 9

Thrashing

Answer 9

The system spends too much time swapping out data and not executing

Question 10

What prevents thrashing?

Answer 10

The OS keeps track of what has been used least recently when moving things around

Question 11

Principle of locality

Answer 11

Program and data references within a process tend to cluster

Question 12

How is a page table modified for virtual memory?

Answer 12

It indicates if a page is present in main memory and if it has been modified

Question 13

How many page tables are there?

Answer 13

One per process

Question 14

How are large page tables handled?

Answer 14

They too can be put into virtual memory. Only parts of the table are in main memory at a time

Question 15

What is an inverted page table?

Answer 15

One that hashes the page number portion of a virtual address. the hash points to the inverted page table

Question 16

What does an inverted page table contain?

Answer 16

The addresses of a real memory page frame in lieu of one per virtual address

Question 17

What makes the inverted page table inverted?

Answer 17

It addresses by frame number instead of virtual page number

Question 18

How many main memory accesses are there per virtual memory access?

Answer 18

Two. One for the page table and one for the data

Question 19

translation look aside buffer (TLB)

Answer 19

Caches functions and contains page table entries that have most recently been used

Question 20

What happens if something is not found in the TLB?

Answer 20

It indexes the process page table, and gets the data, then updates the TLB

Question 21

Associative mapping

Answer 21

The processor stores the entire page table entry in the TLB and checks the TLB entries when it’s looking for a page table entry

Question 22

What happens if pages are too small?

Answer 22

There are too many pages and more faults

Question 23

What happens if a page is too large?

Answer 23

They begin to contain data further and further away from what is needed, which violates locality and causes more faults

Question 24

What happens if a program jumps around? (Object oriented)

Answer 24

There are fewer TLB hits and locality decreases. The TLB actually causes bottlenecking

Question 25

Why do most OSes only support one page size?

Answer 25

Because page sized effect OS performance throughout

Question 26

What is the difference between segmentation and paging?

Answer 26

Paging is invisible to the programmer

Question 27

What is a ring structure?

Answer 27

A method of assigning program privileges. Kernel functions get ring 0

Question 28

What 3 things are considered in OS memory management?

Answer 28

Using virtual memory, paging/segmentation/both, and management algorithms

Question 29

What is fetch policy?

Answer 29

The means by which it is decided if a page should be brought into main memory

Question 30

Demand paging

Answer 30

Only binging a page into memory when a request is made for that page. Lots of faults at first, then locality kicks in

Question 31

Prepaging

Answer 31

Bringing in a lot of contiguous pages at once. This fails if the ones brought in are never referenced

Question 32

Placement Policy

Answer 32

Determining where in main memory a process should reside. Usually handled with hardware rather than software

Question 33

Replacement Policy

Answer 33

Determining how to replace a page when a new page must be brought in

Question 34

Resident Set Management

Answer 34

How many page frames are to be allocated per process, and if pages replaced should be limited to the process causing a fault

Question 35

Frame Locking

Answer 35

Frames that cannot be replaced. Usually applies to kernel/os stuff

Question 36

Optimal replacement policy

Answer 36

Replace the page with the longest time to the next reference. Has least number of page faults

Question 37

Least recently used replacement policy

Answer 37

Replaces the page that has the longest time since the last reference. Locality says this is least likely to be referenced in the future.

Question 38

First in first out replacement policy

Answer 38

Page frames are treated as a circular buffer. Simple to implement but performs poorly.

Question 39

Clock policy

Answer 39

Tries to implement LRU in a fashion with little overhead. Blends LRU with FIFO

Question 40

How does clock policy work?

Answer 40

Each page frame has a use bit. When it’s time to replace, it replaces with the first thing that has a 0 use bit. Anything looped through until that zero has a use bit set to 0.