1. Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Virtual Memory

2. 9.2 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition What to Learn? Chapter 8 in the text book OSCE. Chapter 9 in OSC. The benefits of a virtual memory system The concepts of demand paging page-replacement algorithms and allocation of physical page frames Other related techniques Copy-on-write. Memory mapped files

3. 9.3 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Background Virtual memory Only part of the program needs to be in memory for execution Logical address space can be much larger than physical address space Allows address spaces to be shared by several processes

4. 9.4 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Virtual Memory That is Larger Than Physical Memory

5. 9.5 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Transfer of a Paged Memory to Contiguous Disk Space

6. 9.6 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Demand Paging for Virtual Memory Bring a page into memory ONLY when it is needed Less I/O needed Less memory needed Faster response More users supported Check a page table entry when a page is needed not-in-memory bring to memory

7. 9.7 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Valid-Invalid Bit With each page table entry a valid–invalid bit is associated (v in-memory, i not-in-memory) Initially valid–invalid bit is set to i on all entries Not in memory page fault v v v v i i i …. Frame #valid-invalid bit page table

8. 9.8 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Page Table When Some Pages Are Not in Main Memory

9. 9.9 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Handling a Page Fault 1. Get an empty physical frame 2. Swap page from the disk into frame 3. Reset the page table Set validation bit = v 4. Restart the instruction that caused the page fault

10. 9.10 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Page Fault (Cont.) Restart instruction Special handling block move Auto increment/decrement

11. 9.11 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Steps in Handling a Page Fault

12. 9.12 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Performance of Demand Paging Page Fault Rate 0 p 1.0 if p = 0, no page faults if p = 1, every reference is a fault Effective Access Time (EAT) EAT = (1 – p) x memory access + p (page fault overhead + swap page out + swap page in + restart overhead )

13. 9.13 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Demand Paging Performance Example Memory access time = 200 nanoseconds Average page-fault service time = 8 milliseconds EAT = (1 – p) x 200 + p (8 milliseconds) = (1 – p x 200 + p x 8,000,000 = 200 + p x 7,999,800 If one access out of 1,000 causes a page fault, then EAT = 8.2 microseconds. This is a slowdown by a factor of 40!!

14. 9.14 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Other Benefits Virtual memory allows other benefits - Shared pages - Copy-on-Write during process creation - Memory-Mapped Files

15. 9.15 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Shared Pages with Virtual Memory

16. 9.16 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Copy-on-Write: Lazy copy during process creation Copy-on-Write (COW) allows both parent and child processes to initially share the same pages in memory If either process modifies a shared page, this page is copied COW allows more efficient process creation as only modified pages are copied

17. 9.17 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Before Process 1 Modifies Page C

18. 9.18 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Copy on Write: After Process 1 Modifies Page C

19. 9.19 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Memory-Mapped Files Allow file I/O to be treated as routine memory access by mapping a disk block to a page in memory Simplifies file access by treating file I/O through direct memory access rather than read() write() system calls File access is managed with demand paging. Several processes may map the same file into memory as shared data

20. 9.20 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Memory Mapped Files

21. 9.21 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Example of Linux mmap #define NUMINTS 1000 #define FILESIZE NUMINTS * sizeof(int) fd = open(FILEPATH, O_RDWR | O_CREAT | O_TRUNC, 0600); result = lseek(fd, FILESIZE-1, SEEK_SET); result = write(fd, "", 1); map = mmap(0, FILESIZE, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); for (i = 1; i <=NUMINTS; ++i) map[i] = 2 * i; munmap(map, FILESIZE); close(fd);

22. 9.22 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Demand paging when there is no free frame? Page replacement – find some page in memory, but not really in use, swap it out Algorithm Performance – want an algorithm which will result in minimum number of page faults Same page may be brought into memory several times

23. 9.23 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Need For Page Replacement

24. 9.24 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Page Replacement

25. 9.25 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Basic Page Replacement 1. Find the location of the desired page on disk 2. Find a free frame: - If there is a free frame, use it - If there is no free frame, use a page replacement algorithm to select a victim frame 3. Swap out: Use modify (dirty) bit to reduce overhead of page transfers – only modified pages are written to disk 4. Bring the desired page into the free frame. Update the page and frame tables

26. 9.26 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Expected behavior: # of Page Faults vs. # of Physical Frames

27. 9.27 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition First-In-First-Out (FIFO) Algorithm Reference string: 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5 3 frames (3 pages can be in memory at a time per process). Maintain a queue and select based on FIFO order. 4 frames Beladys Anomaly: more frames more page faults 1 2 3 1 2 3 4 1 2 5 3 4 9 page faults 1 2 3 1 2 3 5 1 2 4 5 10 page faults 4 43

28. 9.28 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition FIFO Page Replacement

29. 9.29 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition FIFO Illustrating Beladys Anomaly

30. 9.30 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Optimal Algorithm Replace page that will not be used for longest period of time 4 frames example 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5 How do you know this? Used for measuring how well your algorithm performs 1 2 3 4 6 page faults 4 5

31. 9.31 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Optimal Page Replacement

32. 9.32 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Least Recently Used (LRU) Algorithm Reference string: 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5 Counter implementation Every page entry has a counter; every time page is referenced through this entry, copy the clock into the counter When a page needs to be changed, look at the counters to determine which are to change 5 2 4 3 1 2 3 4 1 2 5 4 1 2 5 3 1 2 4 3

33. 9.33 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition LRU Page Replacement

34. 9.34 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition LRU Algorithm Implementation Stack implementation – keep a stack of page numbers in a double link form: Page referenced: move it to the top requires 6 pointers to be changed No search for replacement

35. 9.35 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Use Of A Stack to Record The Most Recent Page References

36. 9.36 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition LRU Approximation Algorithms Reference bit With each page associate a bit, initially = 0 When page is referenced bit set to 1 Replace the one which is 0 (if one exists) We do not know the order, however Second chance If page has reference bit = 1 then: Leave this page in memory, but set reference bit 0. Visit next page (in clock order) If reference bit is 0, replace it.

37. 9.37 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Second-Chance Page-Replacement Algorithm

38. 9.38 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Issues with Page Allocation and Replacement Initial Allocation. Each process needs minimum number of pages Two schemes: Fixed allocation vs. priority allocation Where to find frames Global replacement – find a frame from all processes. Local replacement – find only from its own allocated frames

39. 9.39 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Fixed Allocation Equal allocation – For example, if there are 100 frames and 5 processes, give each process 20 frames. Proportional allocation – Allocate according to the size of process

40. 9.40 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Priority Allocation Use a proportional allocation scheme using priorities rather than size If process P i generates a page fault, select for replacement one of its frames select for replacement a frame from a process with lower priority number

41. 9.41 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Thrashing If a process does not have enough pages, the page-fault rate is very high. This leads to: low CPU utilization operating system thinks that it needs to increase the degree of multiprogramming another process added to the system Thrashing a process is busy swapping pages in and out

42. 9.42 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Thrashing (Cont.)

43. 9.43 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Relationship of Demand Paging and Thrashing Why does demand paging work? Locality model Process migrates from one locality to another Localities may overlap Need a minimum working set to be effective Why does thrashing occur? working-sets > total memory size

44. 9.44 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Locality In A Memory-Reference Pattern

45. 9.45 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Working Sets and Page Fault Rates

46. 9.46 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Prefetching Pages (Prepaging) Optimization to reduce page faults (e.g. during process startup) Preload some pages before they are referenced But if prefetched pages are unused, I/O and memory was wasted Assume s pages are prepaged and α % of the pages is used Page fault cost of s * α > cost of prepaging s * (1- α) unnecessary pages

47. 9.47 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Tradeoffs: Page Size Impact of page size selection: Internal fragmentation Page table size I/O overhead locality

48. 9.48 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Other Issues – TLB Reach TLB Reach - The amount of memory accessible from the TLB TLB Reach = (TLB Size) X (Page Size) Ideally, the working set of each process is stored in the TLB Increase the Page Size Increase TLB reach But may increase internal fragmentation as not all applications require a large page size

49. 9.49 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Other Issues – Impact of Program Structure on Page Faults <128 physical memory pages for each process. Each page is of 128x4 bytes Program structure l Int[128,128] data; Each row is stored in one page Page faults of Program 1? for (j = 0; j <128; j++) for (i = 0; i < 128; i++) data[i,j] = 0; Page faults of Program 2 ? for (i = 0; i < 128; i++) for (j = 0; j < 128; j++) data[i,j] = 0;

50. 9.50 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Other Issues – Impact of Program Structure <128 physical pages for each process. Each page is of 128x4 bytes Program structure l Int[128,128] data; Each row is stored in one page Program 1 for (j = 0; j <128; j++) for (i = 0; i < 128; i++) data[i,j] = 0; 128 x 128 = 16,384 page faults Program 2 for (i = 0; i < 128; i++) for (j = 0; j < 128; j++) data[i,j] = 0; 128 page faults

51. 9.51 Silberschatz, Galvin and Gagne ©2009 Operating System Concepts – 8 th Edition Other Issues – I/O interlock I/O Interlock – Pages must sometimes be locked into memory Consider I/O - Pages that are used for copying a file from a device must be locked from being selected for eviction by a page replacement algorithm