Memory Management in OS
- 1. Memory Management inMemory Management in Operating SystemsOperating Systems V.V. SubrahmanyamV.V. Subrahmanyam SOCIS, IGNOUSOCIS, IGNOU Date: 31Date: 31stst March, 2008March, 2008 Time: 19-00 to 19-45Time: 19-00 to 19-45
- 2. MemoryMemory Memory is central to the operation of a modernMemory is central to the operation of a modern computer system.computer system. Memory is a large array of words or bytes, eachMemory is a large array of words or bytes, each location with its own address.location with its own address. Interaction is achieved through a sequence ofInteraction is achieved through a sequence of reads/writes of specific memory address.reads/writes of specific memory address. The CPU fetches the program from the hard diskThe CPU fetches the program from the hard disk and stores in the memory.and stores in the memory. If a program is to be executed, it must beIf a program is to be executed, it must be mapped to absolute addresses and loaded intomapped to absolute addresses and loaded into memory.memory.
- 3. Contd…Contd… In a multiprogramming environment, in order toIn a multiprogramming environment, in order to improve both the CPU utilisation and the speedimprove both the CPU utilisation and the speed of the computer’s response, several processesof the computer’s response, several processes must be kept in memory.must be kept in memory. There are many different algorithms dependingThere are many different algorithms depending on the particular situation to manage theon the particular situation to manage the memory.memory. Selection of a memory management scheme forSelection of a memory management scheme for a specific system depends upon many factors,a specific system depends upon many factors, but especially upon the hardware design of thebut especially upon the hardware design of the system.system. Each algorithm requires its own hardwareEach algorithm requires its own hardware support.support.
- 4. Memory Management –Memory Management – Responsibilities of OSResponsibilities of OS Keep track of which parts of memory areKeep track of which parts of memory are currently being used and by whom.currently being used and by whom. Decide which processes are to be loadedDecide which processes are to be loaded into memory when memory spaceinto memory when memory space becomes available.becomes available. Allocate and deallocate memory space asAllocate and deallocate memory space as needed.needed.
- 5. Contd…Contd… In the multiprogramming environmentIn the multiprogramming environment operating system dynamically allocatesoperating system dynamically allocates memory to multiple processes.memory to multiple processes. Thus memory plays a significant role inThus memory plays a significant role in the important aspects of computer systemthe important aspects of computer system like performance, S/W support, reliabilitylike performance, S/W support, reliability and stability.and stability.
- 6. OverlaysOverlays In idealistic situation, the entire program and itsIn idealistic situation, the entire program and its related data is loaded in physical memory forrelated data is loaded in physical memory for execution.execution. But what if process is larger than the amount ofBut what if process is larger than the amount of memory allocated to it?memory allocated to it? We can overcome this problem by adopting aWe can overcome this problem by adopting a technique called astechnique called as OverlaysOverlays. Like dynamic. Like dynamic loading, overlays can also be implemented byloading, overlays can also be implemented by users without OS support.users without OS support.
- 7. The entire program or application is divided intoThe entire program or application is divided into instructions and data sets such that when oneinstructions and data sets such that when one instruction set is needed it is loaded in memoryinstruction set is needed it is loaded in memory and after its execution is over, the space isand after its execution is over, the space is released.released. As and when requirement for other instructionAs and when requirement for other instruction arises it is loaded into space that was releasedarises it is loaded into space that was released previously by the instructions that are no longerpreviously by the instructions that are no longer needed.needed. Such instructions can be called as overlays,Such instructions can be called as overlays, which are loaded and unloaded by the program.which are loaded and unloaded by the program.
- 8. DefinitionDefinition: An overlay is a part of an: An overlay is a part of an application, which has been loaded at sameapplication, which has been loaded at same origin where previously some other part (s) oforigin where previously some other part (s) of the program was residing.the program was residing. A program based on overlay scheme mainlyA program based on overlay scheme mainly consists of following:consists of following: A “root” piece which is always memory residentA “root” piece which is always memory resident Set of overlays.Set of overlays. Overlay gives the program a way to extendOverlay gives the program a way to extend limited main storage. An important aspectlimited main storage. An important aspect related to overlays identification in program isrelated to overlays identification in program is concept of mutual exclusion i.e., routines whichconcept of mutual exclusion i.e., routines which do not invoke each other and are not loaded indo not invoke each other and are not loaded in memory simultaneously.memory simultaneously.
- 9. Example of OverlayExample of Overlay Read ( ) Function1 ( ) Function2 ( ) Display ( ) Read ( ) Display ( ) Function2 ( )Function1() 20 K 20K Read( ) Read ( ) 50K Function1( ) Function 1( ) Function 2( ) 70K Function 2( ) Overlay A 40K Display ( ) Overlay B 40K Display ( ) (50K) (70K) Without Overlay (180K) With Overlay (Maximum.130K)
- 10. Limitations of OverlaysLimitations of Overlays Overlays scheme suffers from following limitations:Overlays scheme suffers from following limitations: Require careful and time-consuming planning.Require careful and time-consuming planning. Programmer is responsible for organizingProgrammer is responsible for organizing overlay structure of program with the help of fileoverlay structure of program with the help of file structures etc. and also to ensure that piece ofstructures etc. and also to ensure that piece of code is already loaded when it is called.code is already loaded when it is called. Operating System provides the facility to loadOperating System provides the facility to load files into overlay region.files into overlay region.
- 11. SwappingSwapping Swapping is an approach for memorySwapping is an approach for memory management by bringing each process inmanagement by bringing each process in entirety, running it and then putting it back on theentirety, running it and then putting it back on the disk, so that another program may be loadeddisk, so that another program may be loaded into that space.into that space. Swapping is a technique that lets you use a diskSwapping is a technique that lets you use a disk file as an extension of memory.file as an extension of memory. Lower priority user processes are swapped toLower priority user processes are swapped to backing store (disk) when they are waiting forbacking store (disk) when they are waiting for I/O or some other event like arrival of higherI/O or some other event like arrival of higher priority processes. This ispriority processes. This is Rollout SwappingRollout Swapping.. Swapping the process back into store whenSwapping the process back into store when some event occurs or when needed (may be in asome event occurs or when needed (may be in a different partition) is known asdifferent partition) is known as Roll-in swappingRoll-in swapping..
- 12. SwappingSwapping Operating System User Process/Application ProcessP1 ProcessP2 Main Memory Disk Roll Out Roll In
- 13. Benefits of SwappingBenefits of Swapping Allows higher degree of multiprogramming.Allows higher degree of multiprogramming. Allows dynamic relocation, i.e., if addressAllows dynamic relocation, i.e., if address binding at execution time is being used we canbinding at execution time is being used we can swap in different location else in case of compileswap in different location else in case of compile and load time bindings processes have to beand load time bindings processes have to be moved to same location only.moved to same location only. Better memory utilisation.Better memory utilisation. Less wastage of CPU time on compaction, andLess wastage of CPU time on compaction, and Can easily be applied on priority-basedCan easily be applied on priority-based scheduling algorithms to improve performance.scheduling algorithms to improve performance.
- 14. LimitationsLimitations Entire program must be resident in storeEntire program must be resident in store when it is executing.when it is executing. Also processes with changing memoryAlso processes with changing memory requirements will need to issue systemrequirements will need to issue system calls for requesting and releasing memory.calls for requesting and releasing memory. It is necessary to know exactly how muchIt is necessary to know exactly how much memory a user process is using and alsomemory a user process is using and also that it is blocked or waiting for I/O.that it is blocked or waiting for I/O.
- 15. Logical and Physical AddressLogical and Physical Address SpaceSpace The computer interacts via logical and physicalThe computer interacts via logical and physical addressing to map memory.addressing to map memory. Logical address is the one that is generated byLogical address is the one that is generated by CPU, also referred to as virtual address. TheCPU, also referred to as virtual address. The program perceives this address space.program perceives this address space. Physical address is the actual addressPhysical address is the actual address understood by computer hardware i.e., memoryunderstood by computer hardware i.e., memory unit. Logical to physical address translation isunit. Logical to physical address translation is taken care by the Operating System.taken care by the Operating System.
- 16. Memory AllocationMemory Allocation Memory Allocation Contiguous Allocation Non-Contiguous Allocation Single-partition systems Multiple-partition systems Fixed-Sized Partitions Variable Sized Partitions Equal-Sized Unequal-Size
- 17. PagingPaging In a paged system, logical memory is dividedIn a paged system, logical memory is divided into a number of fixed sizes ‘chunks’ calledinto a number of fixed sizes ‘chunks’ called pagespages.. The physical memory is also predivided intoThe physical memory is also predivided into same fixed sized blocks (as is the size of pages)same fixed sized blocks (as is the size of pages) calledcalled page framespage frames.. The page sizes (also the frame sizes) areThe page sizes (also the frame sizes) are always powers of 2, and vary between 512 bytesalways powers of 2, and vary between 512 bytes to 8192 bytes per page. The reason behind thisto 8192 bytes per page. The reason behind this is implementation of paging mechanism usingis implementation of paging mechanism using page number and page offset.page number and page offset.
- 18. Paging SystemPaging System Pages Page0 Page1 Page2 Page3 Page4 1 3 6 4 2 0 1 2 3 4 Page0 Page4 Page1 Page3 Page2 1 2 3 4 5 6 Frames0 Logical Memory PageNo. FrameNo PageTable Physical Memory
- 19. Page AllocationPage Allocation In variable sized partitioning of memory every time whenIn variable sized partitioning of memory every time when a process of sizea process of size nn is to be loaded, it is important tois to be loaded, it is important to know the best location from the list of available/freeknow the best location from the list of available/free holes. This dynamic storage allocation is necessary toholes. This dynamic storage allocation is necessary to increase efficiency and throughput of system. Mostincrease efficiency and throughput of system. Most commonly used strategies to make such selection are:commonly used strategies to make such selection are: Best-fit Policy:Best-fit Policy: Allocating the hole in which theAllocating the hole in which the process fits most “tightly” i.e., the difference between theprocess fits most “tightly” i.e., the difference between the hole size and the process size is the minimum one.hole size and the process size is the minimum one. First-fit Policy:First-fit Policy: Allocating the first available holeAllocating the first available hole (according to memory order), which is big enough to(according to memory order), which is big enough to accommodate the new process.accommodate the new process. Worst-fit Policy:Worst-fit Policy: Allocating the largest hole that willAllocating the largest hole that will leave maximum amount of unused space i.e., leftoverleave maximum amount of unused space i.e., leftover space is maximum after allocation.space is maximum after allocation.
- 20. Now, question arises which strategy is likely toNow, question arises which strategy is likely to be used?be used? In practice, best-fit and first-fit are better thanIn practice, best-fit and first-fit are better than worst-fit. Both these are efficient in terms of timeworst-fit. Both these are efficient in terms of time and storage requirement.and storage requirement. Best-fit minimize the leftover space, createBest-fit minimize the leftover space, create smallest hole that could be rarely used.smallest hole that could be rarely used. First-fit on the other hand requires leastFirst-fit on the other hand requires least overheads in its implementation because of itsoverheads in its implementation because of its simplicity.simplicity. Possibly worst-fit also sometimes leaves largePossibly worst-fit also sometimes leaves large holes that could further be used toholes that could further be used to accommodate other processes. Thus all theseaccommodate other processes. Thus all these policies have their own merits and demerits.policies have their own merits and demerits.
- 21. Direct MappingDirect Mapping Processor (Running Program) P d f f d Virtual Address Physical Address P Physical Memory PageTableVA → PA
- 22. SegmentationSegmentation Segmentation presents an alternative schemeSegmentation presents an alternative scheme for memory management.for memory management. This schemeThis scheme divides the logical address spacedivides the logical address space into variable length chunks, called segments,into variable length chunks, called segments, with no proper ordering among them. Eachwith no proper ordering among them. Each segment has a name and a length. Forsegment has a name and a length. For simplicity, segments are referred by a segmentsimplicity, segments are referred by a segment number, rather than by a name.number, rather than by a name. Thus, the logical addresses are expressed as aThus, the logical addresses are expressed as a pair of segment number and offset withinpair of segment number and offset within segment.segment.
- 23. It allows a program to be broken down intoIt allows a program to be broken down into logical parts according to the user view oflogical parts according to the user view of the memory, which is then mapped intothe memory, which is then mapped into physical memory.physical memory. Though logical addresses are two-Though logical addresses are two- dimensional but actual physical addressesdimensional but actual physical addresses are still one-dimensional array of bytesare still one-dimensional array of bytes only.only.
- 24. Address TranslationAddress Translation Segment Table Physical Memory Yes S d Limit L BaseB B+d+d<L S Physical AddressNo Trap/ Error dProcessor
- 25. Principle of OperationPrinciple of Operation Segment 0 Segment 1 Segment 2 Segment 3 Segment 4 Limit Base 1000 1500 500 5000 400 3000 1000 4000 1100 6000 1 2 3 4 Segment Table Segment 0 Segment 2 Segment 3 Segment 1 Segment 4 0 1500 2500 3000 3400 4000 5000 5500 6000 7100