Exam 1 Questions

Front 1

When does the CPU interrupt a device?

Back 1

CPU NEVER interrupts a device

Front 2

When does a CPU acknowledge an interrupt?

Back 2

After current instruction execution

Front 3

When does a process switch from the waiting state to the running state?

Back 3

Process NEVER switches from waiting to running

Front 4

What are the reasons for a process to lose CPU in a non-preemptive scheduling system?

Back 4

1. I/O loop
2. terminate process

Front 5

Assume that there are ten processes total currently in a system with five CPUs. Four processes are running, four are in the ready state, and two are in the waiting state.

1. How many PC registers does the system have?

2. How many PC registers are being used by those ten processes?

Back 5

1. 5 PC registers because there are 5 CPUs

2. 4 PC registers are being used (4 are running)

Front 6

Consider a process that has spent 30 units of CPU time, 60 units in the ready state, and 100 units in the I/O waiting queue.

1. What is the turnaround time of the process?

2. What is the waiting time of the process?

Back 6

1. turnaround time = CPU time + ready time = 90 units

2. waiting time = ready state = 60 units

Front 7

What does the PC field in the PCB of a running process indicate?

Back 7

the starting address of the CPU bursts

Front 8

What does the PC field in the PCB of a waiting process indicate?

Back 8

The next instruction to be executed

Front 9

Consider the event of context switching from process Pi to Pj.

Which register will be saved in the PCB of Pi ?

Back 9

PC and registers being used

Front 10

Assume that typical process spends 90% of the time in I/O activity, and the system maintains 20 programs in memory on an average.

1. What is the CPU utilization?

2. What is the probability of the CPU being idle?

Back 10

1. CPU utilization = (1 - Pⁿ) = (1-0.9²⁰)

2. Probability = Pⁿ = 0.9²⁰

Front 11

Consider the following program:

CPU_statement;

Input_statement;

Output_statement;

CPU_statement;

How many CPU and I/O burst would be created by the above program?

Back 11

CPU bursts = # I/O bursts + 1 = 2 + 1 = 3

I/O bursts = 2

Front 12

Consider a system using time slices of 100 ms. Assume process Pi needs 20 seconds of CPU time.

How many times slices would Pi require to complete its task?

Back 12

Pi needs at least 200 time slices because it can take more time

Front 13

Consider a system with time slices of 200 ms. Assume that a process Pi has used a CPU bursts of length 4000ms. Assume that the system has many other processes sharing the CPU with Pi.

How many time slices did Pi use during that CPU bursts?

Back 13

Exactly 20 time slices since you cannot use a time slice in CPU if I/O is not present

Front 14

A CPU scheduler is invoked in many events. List all possible events when a preemptive schedule will be invoked, but a non-preemptive scheduler will not be.

Back 14

- Time out

- submit state to ready state

- waiting state to ready state after I/O done

Front 15

Consider an execution of a set of 20 processes running under the SJN schedule. COnsider another execution of the same set of processes running under the SRT schedule.

In what situation will the total turnaround time be the same for both SJN and SRT schedulers?

Back 15

Processes arrive in non decreasing order of CPU time requirement

Front 16

Consider a system running under a RR scheduler using a time slice of 100 time units. Assume that no process is running in the current state. There are 200 processes in the ready queue and 300 processes in the waiting queues.

How long will the last process in the ready queue wait before getting the CPU?

Back 16

last process will be served in ≤ [(n-1)t == (100 * 199)]

≤ because some processes can use less time slices

Front 17

Consider the Linux scheduler discussed in class. When does a process move from the active array to the expired array?

Back 17

1. There are starving processes in the expired array so the process, even if it is highly interactive, cannot stay in the active

2. If the above is true and the time slice is up

Front 18

Consider the load balancer in the Linux scheduler discussed in class.

1. Which processes have the highest priority to migrate?

2. When can a process be preempted?

Back 18

1. Expired array and not cache hot process have highest priority

2. A process is preempted when its time slice is up or a process joins ready state and priority is higher than the running process

Front 19

Consider the Linux scheduler discussed in class.

When does the kernel check the need_resched flag? Explain the reason.

Back 19

It checks right before it switches from kernel mode to user mode

This is to decrease the amount of switching

Front 20

Give an example of a privileged register.

Back 20

Instruction registers

Front 21

What are the two modes a system run in?

Back 21

Privileged mode and user/non-privileged mode

Front 22

What is a virtual CPU?

Back 22

-time multiplexing, CPU switches multiple programs at high frequencies

-creates illusion of multiple dedicated CPU

Front 23

What is the purpose of multi programming?

Back 23

To increase CPU processes

Front 24

When does a process switch from the running state to ready?

Back 24

After time out

Front 25

What are the reasons for a process to lose a CPU?

Back 25

*State diagram*

running state to ready state

running state to I/O loop

running state to terminated