1. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.1 Operating System Concepts Operating Systems Lecture 38 Frame Allocation Read Ch. 10.5 - 10.6

2. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.2 Operating System Concepts Page Replacement Algorithms Page replacement algorithms select the page to be replaced. Want lowest page-fault rate. Evaluate algorithm by running it on a particular string of memory references (reference string) and computing the number of page faults on that string. In all our examples, the reference string is 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5. Last time:  FIFO algorithm  Optimal algorithm

3. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.3 Operating System Concepts Least Recently Used (LRU) Algorithm Replace the page that has not been used for the longest period of time. Reference string: 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5 Considered a good replacement algorithm. Question: How do we implement it?

4. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.4 Operating System Concepts LRU Counter implementation Add a logical clock or counter that is incremented with each memory reference. When a reference is made to a page, the clock register is copied into the time-of-use field in the page table entry. Replace the page with the smallest time value. This requires a search of the page table to find the page with the lowest clock value. It also requires an extra write to memory (of the clock value) for each memory access.

5. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.5 Operating System Concepts LRU Algorithm, Stack Implementation Keep a stack of page numbers. If a page is referenced, move it to the top of the stack. The LRU page ends up on the bottom of the stack. No search is required to find the LRU page. Because of the need to remove a page from the middle of the stack, implement the stack with a doubly linked list. Example: 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5

6. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.6 Operating System Concepts Use Of A Stack to Record The Most Recent Page References

7. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.7 Operating System Concepts Allocation of Frames Each process has a maximum number of frames allocated--we cannot allocate more frames than there are in memory. Each process needs minimum number of pages.  As the number of frames per process decreases, the fault rate increases.  When a page fault occurs before an instruction is complete, we must restart the instruction.  There must be enough frames to hold all the pages that a single instruction can reference. Example: IBM 370 – 6 pages to handle Storage to Storage MOVE instruction (SS MOVE):  instruction is 6 bytes, might span 2 pages.  2 pages to handle from.  2 pages to handle to. Two major allocation schemes.  fixed allocation  priority allocation

8. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.8 Operating System Concepts Fixed Allocation Equal allocation – e.g., if 100 frames and 5 processes, give each 20 pages. Proportional allocation – Allocate according to the size of process.

9. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.9 Operating System Concepts Priority Allocation Use a proportional allocation scheme using priorities rather than size. If process P i generates a page fault, either  select for replacement one of its frames.  OR select for replacement a frame from a process with lower priority number.

10. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.10 Operating System Concepts Global vs. Local Allocation Global replacement – process selects a replacement frame from the set of all frames; one process can take a frame from another. Local replacement – each process selects from only its own set of allocated frames. In global replacement a process cannot control its own page fault rate. It's turnaround time may vary depending on the behavior of other processes. With local replacement, a process may be hindered because it does not have access to other, less used pages in memory. Global replacement generally has greater system throughput, so it is more commonly used.

11. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.11 Operating System Concepts Thrashing If a process does not have “enough” pages, the page-fault rate is very high. Thrashing  a process is busy swapping pages in and out. Cause of thrashing:  If using a global replacement scheme, if one process starts needing more frames it will take them from other processes.  These other processes may start faulting more.  There could end up being many processes in queue waiting for pager to swap in needed pages.  This leads to low CPU utilization.  operating system thinks that it needs to increase the degree of multiprogramming to increase CPU utilization.  another process added to the system, making the problem worse.

12. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.12 Operating System Concepts Thrashing Why does paging work? Locality model  Process migrates from one locality to another.  Localities may overlap. Why does thrashing occur?  size of locality > total memory size

13. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.13 Operating System Concepts Locality In A Memory-Reference Pattern

14. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.14 Operating System Concepts Working-set model The working-set model is a way of estimating the size of the current locality for a process.   working-set window  a fixed number of page references The working set is the set of pages in the most recent  page references.

15. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.15 Operating System Concepts Working-Set Model WSS i (working set of Process P i ) = total number of pages referenced in the most recent  (varies in time)  if  too small will not encompass entire locality.  if  too large will encompass several localities.  if  =   will encompass entire program. D =  WSS i  total demand frames if D > m  Thrashing Policy if D > m, then suspend one of the processes.

16. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 7.16 Operating System Concepts Page-Fault Frequency Scheme The Page-Fault Frequency scheme is an alternative to the working-set model. Establish “acceptable” page-fault rate.  If actual rate too low, process loses frame.  If actual rate too high, process gains frame.