CHAPTER-8

VIRTUAL MEMORY

 Virtual memory is a technique of executing program instructions that may

not fit entirely in the system memory.
 In this scheme, the programs can be larger than the actual physical memory
available.
 The basic concept of virtual memory is to store instruction and data of a
program in the secondary memory and to load them in the main memory
when they are required.
 Virtual memory thus separates the user logical memory from physical
memory.
 Virtual memory is commonly implemented by demand paging and demand
segmentation.

 Advantages of virtual memory:

1. Unlimited address space: using virtual memory user is able to write
programs for a very large virtual address space.
2. Increased degree of multiprogramming: with the implementation of
virtual memory, each user program occupies less physical memory, as a
result more programs occupies less physical memory, as a result more
programs can be run at the same time.
3. Reduced I/O: less I/O is needed to load or swap each user program into
memory, so each user program will run faster.
4. Automatic memory allocation: memory is allocated automatically on
the basis of demand.
5. Efficient memory utilization: it utilizes memory efficiently as entire
process need not be in the main memory.
 Disadvantages of virtual memory:
1. It requires additional hardware support for address translation.
2. Possibility of thrashing due to excessive paging and page faults.
 DEMAND PAGING:
 Demand paging is a technique used to implement the concept of
virtual memory.
 It combines the concept of paging and swapping.
  Whenever a process is to be executed, it is brought into the main
memory.
 In demand paging, only those pages of the process are brought into
the memory that are required. All other pages reside on secondary
storage and are brought into the memory when needed.
 This task of bringing in and out the various pages is done by pager.
 This concept is called demand paging because the page is not loaded
into the memory until it is not required.

 PURE DEMAND PAGING:

 It is the form of demand paging in which not even single page is
loaded into the memory, initially.
 Thus very first instruction causes a page fault in this case as it is not
residing in main memory.
 This kind of demand paging may significantly decrease the
performance of a computer system.
 PAGE REPLACEMENT ALGORITHMS:
 When a page fault occurs, the operating system has to choose a page
to remove from memory to make space for that page to be brought in.
 If the page to be removed has been modified in memory, it must be
rewritten to the disk to bring the disk copy up-to-date.
  Thus page replacement algorithm is the logic or the policy regarding
how to select a page to be swapped out from main memory to create a
space for the page that has caused a page fault.
 OPTIMAL PAGE REPLACEMENT ALGORITHM:
 This algorithm considered as the best page replacement algorithm in a
virtual memory.
 This algorithm is also called OPT or MIN.
 In this algorithm, we replace a page that will not be used for longer
period of time.
 The page which is not used for longer period of time is swapped out
from the main memory to create space for the requested page.

 Advantages of optimal page replacement:

 It has the lowest rate of occurrence of page faults.
 It improves the system performance by reducing overhead for
numbering of page faults and swapping pages in and out, when a page
fault occurs.
 Disadvantages of optimal page replacement:
 It is difficult to implement because it requires future knowledge of the
reference string.
 First-in First out (FIFO) page replacement algorithm:
 It is simplest page replacement algorithm and associates with each
page the time when that page was brought into memory.
 In this algorithm, whenever a page is to be replaced, the oldest page is
chosen it means the page is swapped out which has been in the main
memory for longest period of time.
 It is implemented by creating FIFO queue that holds all pages in
memory.
 The page at the head of the queue is replaced and when a page is
brought into memory, it is inserted at the tail of the queue.
  The main difference between FIFO and OPT algorithm is that FIFO
algorithm uses the time when a page was brought into the memory.
The OPT algorithm uses the time when a page is to be used.
 ADVANTAGES OF FIFO PAGE REPLACEMENT:
 It is simple and easy to implement.
 It can be easily coded.
 DISADVANTAGES OF FIFO PAGE REPLACEMENT:
 It replaces heavily used pages.
 It suffers from belady’s anomaly.
 Belady’s anomaly:
 When number of page frames allocated to a process is increased, the
number of page faults also increases.

 Least recently used (LRU) page replacement algorithm:

  This algorithm uses information about the pages accessed in recent
past to predict near future.
 In this algorithm, whenever a page is to be replaced, we select the
page that has not been used for the longest period of time.
 This algorithm keeps track of the last time each page was used.
 Thus, LRU algorithm is based on the strategy of looking backward in
time.

 Implementation of LRU page replacement:

1. Using Counters:
 In this method a counter or clock is used.
 Whenever a reference to a particular page is made, its counter is
incremented.
 The LRU algorithm searches the page table for entry with lowest
counter value and selects that page for replacement.
2. Using Matrix:
 On a system with n page frames, an n*n matrix can be used to
implement LRU algorithm.
 The matrix is initialized to contain all 0’s.
 When a page frame k is accessed, all bits in row k are set to 1.
 Then all bits in column k are set to 0.
 At any instant, the row whose binary value is least recently used,
the row whose value is next lowest is next least recently used and
so on.
3. Using Linked list:
 Whenever a page frame is referenced, the entry for that page is
placed at the head of the list.
 Older entries move towards tail of the list.
  Whenever a page is to be replaced, the entry at the tail of the list is
selected for replacement.
 Advantages of LRU page replacement:
 It does not suffer from belady’s anomaly.
 It is less complicated.
 Disadvantages of LRU page replacement:
 It requires expensive hardware.
 It requires additional data structure such as linked list to implement.
 Least Frequently used (LFU) page replacement algorithm:
 It is the approximation of LRU algorithm.
 A counter is used to count the number of references that have been
made to each page.
 Whenever a page is to be replaced, that page is chosen that is least
frequently used or least intensively referenced.
 Most frequently used (MFU) page replacement algorithm:
 it is the reverse of LFU.
 This algorithm is also an approximation of LRU and uses a counter to
count the number of references made to each page.
 Whenever a page to be replaced that page is chosen which has been
used most frequently.
 Not frequently used (NFU) page replacement algorithm:
 NRU is an approximation of LRU.
 The page in the lowest number group should be replaced first and
those in the highest numbered groups should be replaced last.
 Pages within a group are selected randomly for replacement.
 Second chance page replacement algorithm:
 It is the combination of LIFO and NRU algorithm.
 In this algorithm, the page present for the longest time in the memory
is given second chance to remain loaded in the main memory.
 When a page is selected, its reference bit is checked.
 Clock page replacement algorithm:
 it is the variation of second chance page replacement algorithm.
 It uses a circular queue to contain all the pages in the form of a clock.
 It make use of a pointer to indicate the position of the page to be
replaced.
 LOCALITY:
 Locality can be exhibited in two forms:
1. Temporal locality
2. Spatial locality
 Temporal locality: it means the storage locations referenced recently
are likely to be referenced in the near future.
 Example: if weather is sunny at 3.PM, then there is a good
chance that weather was sunny at 2.30PM and will be sunny at
3.30PM.
 Spatial locality: it means that storage references tends to be clustered
so that once a location is referenced, it is highly likely that nearly
locations will be referenced.
 Example: if it is sunny in one town, then it is likely to be sunny
in nearby town.
 ALLOCATION ALGORITHM:
There are two different methods of allocating frames:
1. Equal allocation
2. Proportional allocation
 Equal allocation: this allocation method divides the available page
frames equally among all the processes
Example: there are 93 frames and 5 processes. We can divide 18
frames to each of the 5 processes. The remaining 3 frames are used as
free frame buffer pool.
  Proportional allocation: it divides the frame in proportion to the size
of the process. It means memory is allocated to each process
according to its size.
 Global versus local allocation:
1. Global replacement: in this, the replacement frame selected by a
process can be any frame from the set of all frames.
 In this strategy, one process can take frames from another process.
2. Local replacement: in this, the replacement frame selected by a process
can only be from the set of frames which are allocated to it.
 the process cannot select the frame which is currently allocated to
other process.