Implementation of modern memory management system , Operating System

Central to implementation of a modern memory management system is the page replacement algorithm. Modern virtual memory systems break memory up into pages and map (via a page table) only the pages needed, or a maximum number of pages that is constrained by system load, into the address space of the process. The page replacement algorithm is used to decide which page/s of memory to evict when different pages of memory are needed.

Using the simulators provided you will develop an understanding of the algorithms involved and perform a comparison of the efficiency of these algorithms.

2. The Simulators

The speci?cations of the simulators are as follows:

  • Only a single process run is simulated, so all the available physical memory constitutes the resident set for this process. This represents ?xed allocation, local scope policy for the resident sets.
  • The physical memory size (in frames) is an input parameter to the program. This should be less than the number of virtual pages, so some of the process pages will not ?t into physical memory.
  • The virtual address is 16-bit long. The page is addressed by the high-order 8 bits, and the offset within the page by the low-order 8 bits, so there are 256 virtual pages, each capable of holding 256 bytes.

The following data structures are provided:

  • A linear page table, which maps the virtual page number to a physical frame number. Each page table entry contains the physical frame number, and a valid bit. Other control bits or items should be added to facilitate implementation of the chosen page replacement strategy.
  • A free list, which is a list of all frames in physical memory available for allocation to the process. Initially all available frames belong to this list.
  • A resident set, which is a list of physical frames allocated to the process. Initially  this list is empty.

Initially all available frames are free, and can be allocated to the process on demand. When the free list runs out of frames, one of the frames in the resident set must be freed for the incoming page. The choice of the victim page is governed by the replacement policy.

The execution of the process is simulated by the pager program which reads virtual addresses from the input ?le, and allocates new frames to the process as required.

3 Simulator Algorithms

Three simulators (that share a common codebase) have been provided. They implement the following algorithms:

  • FIFO
  • LRU
  • Optimal

You should research each of these algorithms and develop an understanding of how these algorithms work. You do not need to understand the code beyond the level of how to run the simulators.

4 Algorithm Comparison

Four input ?les have been provided that simulate different process workloads. You should use these as input to the simulators, varying the memory size to determine which is the most ef?cient algorithm for each input, and which is the most ef?cient overall. Plotting available memory versus page fault rate will help you in this task. A spreadsheet has also been provided to assist you in making these plots.

You should after performing these experiments be able to make a recommendation about which algorithm is best for use in a real system.

5 Deliverable and Marking Guide

You are required to write a report that includes the following information, with the marks available are indicated on the right hand side:

  • Introduction to the report
  • Description of "FIFO"
  • Description of "LRU"
  • Description of "Optimal"
  • Description of experiments performed, this should be an overview of the techniques used, the results will pin this down.
  • Presentation of results for input ?le 1, (this could just be a graph), and a ranking of the three algorithms performance.
  • Presentation of results for input ?le 2 and a ranking of the three algorithms performance.
  • Presentation of results for input ?le 3 and a ranking of the three algorithms performance.
  • Presentation of results for input ?le 4 and a ranking of the three algorithms performance.
  • Conclusions and recommendation of algorithm, this should be supported by your data.
Posted Date: 2/25/2013 5:21:13 AM | Location : United States







Related Discussions:- Implementation of modern memory management system , Assignment Help, Ask Question on Implementation of modern memory management system , Get Answer, Expert's Help, Implementation of modern memory management system Discussions

Write discussion on Implementation of modern memory management system
Your posts are moderated
Related Questions

Explain about deadlock prevention? In order for the occurrence of deadlock, the four conditions like mutual exclusion, hold and wait, no pre-emption and  circular wait must hap

Since each thread has its own processing space therefore communication between threads will need to be done through a common global variable. Since multiple threads can access the


What are the drawbacks of contiguous allocation of disk space? The disadvantages are a. Suffers from internal fragmentation b. Suffers from external fragmentation c. F

Define the Global vs. local variables It is also important to discuss the scope of resources being granted to the application program.  Variable scoping is defined as the acces

What is dispatch latency? The time taken by the dispatcher to stop one process and begin another running is known as dispatch latency.

Write on short note DMA. Direct M emory Access (DMA) is a technique for transferring data from  main to a device without passing it through the CPU. Computers that have

What are race conditions? Race conditions are problems that take place due to the sharing of the same file by several processes. In such a case none of the processes is able to

List and discuss the various services provided by the operating system. Program execution - system capability to load a program into memory and to run it. I/O operatio