Scheduling

Question 1

Boot Sequence

Answer 1

(1) CPU loads boot program from ROM, (2) the boot program examines and checks the memory configuration, builds a configuration structure describing the hardware, loads the OS kernel and gives it the configuration structure, (3) initializes OS

Question 2

Initializing the OS

Answer 2

(1) initialize kernel structures, (2) initialize the state of all hardware devices, (3) creates a number of processes to start operation. (4) if there are user programs, those are run, else it enters the idle loop

Question 3

OS wakes up from…

Answer 3

Interrupts, exception, and system calls (traps)

Question 4

Multiprogramming / Concurrency

Answer 4

One process running while one or more on I/O to increase system utilization and throughput

Question 5

Context Switch

Answer 5

Changing from one process to another

Question 6

What are the steps of a context switch?

Answer 6

(1) Process is running, (2) Process is blocked or interrupted, (3) Switch to kernel, (4) OS saves process state to PCB, (5) OS chooses new process to run, (6) OS loads its state from PCB, (7) Process is running

Question 7

When does the kernel run the scheduler?

Answer 7

A process switches from running to blocked, a process is created or terminated. an interrupt occurs

Question 8

Non-preemptive System

Answer 8

Scheduler runs when a process is blocked, not because of an interrupt

Question 9

Preemptive System

Answer 9

OS makes scheduling decisions during interrupts, but requires interrupt to give control back to OS

Question 10

Throughput

Answer 10

number of processes completed during a unit of time

Question 11

Turnaround Time

Answer 11

length of time to run a process from initialization to termination

Question 12

CPU utilization

Answer 12

percentage of time that the CPU is busy

Question 13

Response Time

Answer 13

time between issuing a command and getting a result

Question 14

Waiting Time

Answer 14

Total amount of time that the process is in the ready queue

Question 15

What are the 5 criteria for comparing scheduling algorithms?

Answer 15

Throughput, turnaround time, cpu utilization, response time, and waiting time

Question 16

To maximize throughput you should be…

Answer 16

minimizing overhead and efficiently using the system resources (CPU, I/O)

Question 17

The Ideal CPU Scheduler

Answer 17

Maximizes CPU utilization and throughput, minimizes response time, waiting time, and turnaround time

Question 18

Shell

Answer 18

interactive application-level program that runs other programs on behalf of the user. It performs a sequence of read/evaluate steps and then terminates.

Question 19

Signal

Answer 19

a small message that notifies a process than an event of some type as occurred in the system.

Question 20

2 reasons a signal is sent

Answer 20

(1) kernel detected a system event or (2) process invoked kill function

Question 21

Pending Signal

Answer 21

a sent signal that has not yet been received. There can be at most one of these.

Question 22

Issues with signals

Answer 22

(1) pending signals are blocked, (2) pending signals are not queued, (3) system calls can be interrupted

Question 23

Nonlocal jumps

Answer 23

user level control that transfers control directly from one function to another. Allows for immediate return from a deeply nested function call. (setjmp will save current environment before jumping)

Question 24

Selecting a time slice for round robin

Answer 24

• too small: throughput suffers because too much time is spent on context switching
• too big: waiting time suffers and it degenerates to FCFS

Question 25

SJF

Answer 25

Shortest Job First. I/O bound jobs will get priority over CPU jobs because its less work.

Question 26

Using past to predict the future

Answer 26

The policy becomes /adaptive/. If the process is I/O bound in the past, it is likely to be I/O bound in the future. The scheduler will favor jobs that use relatively little CPU

Question 27

Multi-Level Feedback Queues

Answer 27

Highest priority and shortest jobs are at the top level. The time slices increase exponentially at lower levels. This policy can cause starvation. If the context switch comes from an I/O it gets moved up a level.

Question 28

What kinds of systems want what criteria the most?

Answer 28

Interactive systems want good response time while batch systems are more about turnaround time.