Round Robin Scheduling Question?

[jeffc1]

I'm having a bit of trouble understanding how average wait time is calculated with round robin.

Using the example below

http://i54.tinypic.com/2rhrd49.png

I simply do not know how the wait times of P1 =16 + 12 + 8 + 4 = 40?

What is the formula?

This is for Operating Systems concepts.

[User Name:]

Process 1(P1) runs for 4 slices, then execution is taken away for 16(+16), then it runs for 4 more, after which execution is taken away for 8(+8), itruns, execution is taken away... you get the pattern.

[jeffc1]

Herm.. Pardon my slowness but i'm still not able to visualize or see what exactly is happening.
Care to elaborate a bit more with examples?

[jeffc1]

Iv'e found many examples on the web, who also do not explain how these numbers were obtained.

[User Name:]

It's just the sum of all the time slices given to other processes. For P1, it would be the sum of all the slices given to P2, P3, P4, and P5. For Pn, it would be the sum of time slices given to all the processes except process n itself.

Scheduling Algorithms - OSDev Wiki

[astrilaa]

Look at the picture and count.
When process P1 is waiting the other processes are running. Other processes are running
in times :
from 4 to 20 = 16
from 24 to 36 = 12
from 40 to 48 = 8
from 52 to 56 = 4
The total waiting time of process P1 is the sum of waiting periods 16 + 12 + 8 + 4 = 40

Analogically count waiting times for other processes.

[brewbuck]

Each process is given a division of time. The divisions are:

P1 gets 20
P2 gets 12
P3 gets 8
P4 gets 16
P5 gets 4

These processes are scheduled around and around in the same order. So every 20+12+8+16+4 time units, each process has executed once. Suppose you wanted to know the time that P1 was NOT executing. That would be the total time of each cycle (20+12+8+16+4) minus the time that P1 WAS executing (20), which is 16+12+8+4.

The average wait time of any process is the weighted average of the wait times of each process, since each process executed for a different amount of time.

[jeffc1]

"Suppose you wanted to know the time that P1 was NOT executing. "

I see how that works for P1.. but P2 would be 60-12 48. The example says P2 is 32.

"Look at the picture and count." I guess its that simple!

[jeffc1]

If anyone has good examples I can use to practice i would appreciate it.

What happens if the process runtime is less than the time quantum? Does the next process run immediately or wait until the time quantum slot is up?

[SlyMaelstrom]

"Suppose you wanted to know the time that P1 was NOT executing. "

I see how that works for P1.. but P2 would be 60-12 48. The example says P2 is 32.

"Look at the picture and count." I guess its that simple!

The reason it's 32 and not 48 is because P2 completes before the 60 intervals being depicted. P2 completes at 44 intervals.

44 - 12 = 32

What the image says is correct about all wait times. Let me draw you a more verbose set of images for your example to help you see it.

[jeffc1]

Thanks < I appreciate the time you guys took to help me out. Its one of those things that are so simple, certain texts dont even bother to explain, yet can mean the difference between getting the answer right or complexly wrong.