Contents:

Overview

Internet Requirements

Practice Problems

Problems to be Submitted (20 points)

1. (3 points)

Exercise 70 of Ch. 10 (p. 360)

You are asked to draw a Gantt chart. If you cannot easily mimic the diagrams from the text, please feel free to report the same information in whatever way you choose. For example, Exercise 71 of the text is a practice problem with the answer in the back of the text. I might choose to typeset that same answer in a table form, such as:

From     To     CPU Runs
------------------------
  0       50       p4
 50      110       p2
110      230       p1
230      410       p3
410      710       p5

2. (6 points)

Exercise 72 of Ch. 10 (p. 360)
3. (4 points)

If, in a dynamic-partition memory management system, the current value of the base register is 58712 in decimal, and the current value of the bounds register is 1587, compute the physical addresses that correspond to the following logical addresses:
1. 1422
2. 2535
4. (3 points)

Assume that dynamic partitioning is used for memory management and that currently there are six empty blocks with respective sizes 38, 85, 51, 99, 77, 27. A new job arrives requiring 52 blocks of main memory. Which size block would be used be each of the following partition selection approaches:
1. first fit
2. best fit
3. worst fit
5. (4 points)

In a paged memory-management system, the frame and page sizes are 2048 and the following page map table applies to the currently executing process,

Page	0	1	2	3	4
Frame	64	22	8	71	55

Compute the physical addresses that correspond to the following logical addresses:
1. 4100
2. 11000

Extra Credit (2 points)

Consider the following situation:

Job	Arrival Time	Service Time
p1	0	40
p2	0	20
p3	10	5
p4	11	1

As described in the book, the Shortest Job Next (SJN) CPU scheduling algorithm is non-preemptive. Therefore, it produces the following Gantt chart (for simplicity, I'm typesetting the "chart" as the following table):
 

from	to	CPU runs
0	20	p2
20	21	p4
21	26	p3
26	60	p1

Notice that at time 0, processes p3 and p4 have not yet arrived, and so of those processes which are ready, p2 is the shortest. It runs from time 0 to time 20. At time 20, when the scheduler decides on the next job, p1, p3, and p4 are available, and so p4 is the next to run as it is the shortest. After that job finishes, then p3 runs, and finally p1.

What is interesting to note is that SJN is non-preemptive. That is, at time 10, when job p3 arrives, p3 is even shorter than the job which is currently running. However, the SJN scheduling algorithm lets p2 complete. Similarly when p4 arrives.

A common preemptive variant of this is one which is called Shorest Remaining Processing Time (SRPT). With this scheme, if a new job arrives with a service time even less than the currently running job's remaining processing time, then the current job is preempted in favor of the new job. When any job completes, the scheduler will next give the CPU to the ready job with the smallest remaining processing time. So on the above example, the SRPT algorithm produces the following schedule:

from	to	CPU runs
0	10	p2
10	11	p3
11	12	p4
12	16	p3
16	26	p2
26	60	p1

Your extra credit task is the following.
Please give the Gantt chart which results from running the SRPT scheduling algorithm on the following set of jobs:

Job	Arrival Time	Service Time
p1	0	15
p2	0	20
p3	10	1
p4	12	13
p5	14	5
p6	24	4

Solution to the text's problem 35:  Though the physical machine has many users trading demands on it, users see what appear to be their own individual machines, their virtual machines, with the real sharing and interleaving of demands hidden.
Solution to the last practice problem:
   logical address 85 = 0*1024 + 85.  Physical address 5*1024 + 85 =  5205
   logical address 2359 = 2*1024 + 311.  Physical address 7*1024 + 311 = 7479
   logical address 6000 = 5*1024 + 880.  Error:  out of bounds; no page number 5