COMP 2432 Operating Systems Written Assignment Submission to BlackBoard Deadline: 31 March 2021

Draw Gantt charts for the following set of processes using different scheduling algorithms: (a) SRT, (b) Priority with preemption (Linux convention), (c) Priority with preemption (Windows convention),

(d) RR with quantum Q = 3, and (e) RR with quantum Q = 2. Let us assume the standard tie-breaking rules when multiple processes are eligible, i.e., the process with smaller ID will go first. Compute the waiting time and turnaround time for each process under the algorithms (a) to (e).

Process Burst Time Arrival Time Priority P1 9 0 4

P2 5 1 3

P3 1 2 1

P4 9 4 2 P5 7 5 5

(f) Response time measures the time taken for the first response of a process. However, response time does not always provide a good measure for the responsiveness of interactive processes since it only measures the time taken for the first response. Waiting time is a useful measure for system performance too, but normally we just measure the total waiting time for a process by adding up all the waiting periods. The first waiting period is normally more important, as it measure the response time at the beginning of a process. As execution continues, subsequent waiting periods are not as much as a concern. Here, let us take a weight on each waiting period and add them up, instead of simply adding them up in our normal total waiting time computation. In summary, assume that a weighting factor w is adopted. The first waiting period will take a weight of 1, the second period a weight of w, the third period a weight of w2, the fourth period a weight of w3, and so on. The resulting waiting time would be the total weighted waiting time. If w = 1, this is the same as the total waiting time, and if w < 1, this will be smaller. Compute this total weighted waiting time for the processes for w = 0.9 and w = 0.8. Compute also the total weighted waiting time for the extreme case of RR with quantum Q = 1. Note that for non-preemptive algorithms, total weighted waiting time is the same as total waiting time as there is only one waiting period. State briefly your observations.

2. Multi-level Scheduling.

A multi-level feedback queue is adopted to schedule the following processes for execution. Processes first enter the high priority queue based on RR with quantum 2. A process that cannot complete its execution

after a service time of 4 will be demoted to the medium priority queue based on RR with quantum 3. A process that cannot complete its execution after a service time of 6 in this queue will be demoted to the low priority queue based on FCFS.

(a) Assume that Fixed Priority scheduling is adopted. Draw a Gantt chart for the process execution schedule. Compute the process waiting and turnaround time.

(b) Assume that Time Slicing scheduling is adopted with a ratio of 5:3:2 in allocated time slices, in the format of 10 time units, 6 time units and 4 time units respectively on the three queues. Draw a Gantt chart for the process execution schedule. Compute the process waiting and turnaround time.

Process Burst Time Arrival Time Priority P1 15 0 4

P2 7 1 3

P3 11 2 1

P4 5 4 2 P5 12 6 5

3. Contiguous Memory Allocation.

A memory system currently has three holes of size 210K, 291K, and 254K and the total memory is 1000K. You are given 12 requests: 84, 90, 97, 81, 79, 20, 77, 64, 56, 68, 22, and 38K, arriving in that order. Indicate how the requests are satisfied with (a) first-fit algorithm, (b) best-fit algorithm, (c) worst- fit algorithm, and (d) an optimal algorithm that makes a decision after seeing all requests using an oracle. There are multiple possible answers for the optimal algorithm. You are to give two possible answers to (d). What are the utilizations for the four algorithms?

(e) It is found that there are some small off-by-one errors in managing the holes and keeping track of the sizes. The total size is found to increase from 755K to 756K, since the size of one of the three holes has been under-reported by 1K. Since it is unknown which hole has been under-reported for its size, you would try three different cases (i) to (iii) to see whether an improvement to the optimal algorithm can be made: (i) 211K, 291K, 254K, (ii) 210K, 292K, 254K, (iii) 210K, 291K, 255K. Give the request satisfaction scenario if utilization can be improved in some of the three. Otherwise, please state that no improvement is possible.

(f) You are given some tasks of size xK, yK and zK respectively and there are at most 9 tasks each. Determine the best that you could achieve if you are to fill up a hole of size SK, for the following values of x, y, z and S, by minimizing the unused space left, if any. Show how many tasks of size x, how many of size y and how many of size z are used in each case. (g) Now if we relax the requirement so that there are no upper limit on the number of tasks for each type, determine the minimal unused space left and the task mix. (h) If we tighten the requirement so that there are still at most 9 tasks of each size, but we also require that each type of tasks must be used at least once, determine the minimal unused space left and the task mix.

x 36 26 29 y 49 57 39 z 59 62 49 S 800 775 570

4. Segmentation.

Consider the segment table for process P1 containing the following.

P1 Segment Base Length / Limit

0 3011 1 1901 2 5678

4 4434 5 1011 6 3901

135 234 543

787 345 135

3 |
2432 |
304 |

Suppose that segment 3 of P1 is a shared segment with segment 6 of P2, and P2 has 7 segments of size 55, 604, 103, 212, 72, 352 and 304 respectively, starting from segment 0 to segment 6. Note that segment 3 of P1 has the same size as segment 6 of P2, since it is shared. Assume that the computer has a memory of 8KB allocated for users, with physical address from 0 to 8191. Those bytes at the lowest end of the memory as well as those at the highest end of the memory, from 0 to 999, and from 6789 to the end, are reserved by the operating system. Segments for P2 are to be allocated from the free memory in the order of 0, 1, 2, 3, 4, 5, and 6, if necessary, using (a) first-fit algorithm, (b) best-fit algorithm, (c) worst-fit algorithm. Show the three segment tables for P2 for segments allocated under the three algorithms. Note that it is useful to draw diagrams showing the memory allocation to the used segments for clarity.

(d) Translate the following logical addresses for P1 and P2 by filling in the table below.

Allocation algorithm for P2

FF BF WF

Physical address for P2

Logical address (0, 44)

(1, 231)

(2, 82)

(3, 199)

(4, 56)

(5, 304)

(6, 135)

Physical address for P1