CS 3733 Operating Systems, Fall 2002 Assignment 2 Comments
The assignment was graded on a basis of 30 points.
Parts 0 and 1 were worth 10 points together.
You got 10 points for just handing in Part 1.
It was not checked.
Parts 2 and 3 were 10 points each and I looked carefully at the answers you
gave to the questions.
Part 2
- The load average uses the total waiting time, not the average.
- Look at the numbers you calculate to see of they make sense.
If you calculated a load average of .3, most of the time the ready queue
is empty and you would expect a low CPU utilization.
- No units are given for the simulation.
Do not assume they are in microseconds or milliseconds.
- Make sure your Gantt charts are labeled appropriately.
- There is not much difference between a CPU utilization of 100% and one of 98%
Part 3
- What is meant by a large quantum or a small quantum?
These only make sense in comparison to something, like the CPU burst time.
- You need to show tables, not just the graphs.
The graphs do not tell you about average waiting time.
- You cannot prove that things are always the same using a simulation.
- You can show that things are not always the same using a simulation.
- Don't use a CPU burst longer than the duration.
- Don't make your experiments too simple, with one burst per process
or a very small number of processes.