Scribd
Upload
36 views

Unit II: Performance Comparison

This document discusses different methods for comparing the performance of CPU scheduling algorithms: deterministic modeling, queuing analysis, and simulators. Deterministic modeling uses a predetermined workload to directly compare algorithms, but only applies to that specific scenario. Queuing analysis uses formulas like Little's Law to calculate metrics like wait time given arrival and service rates. Simulators provide the most accurate evaluation by programming a computer system model and gathering statistics from executing different algorithms on sample processes and workloads.

Uploaded by

Indian Best
Copyright
© © All Rights Reserved
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
36 views

Unit II: Performance Comparison

This document discusses different methods for comparing the performance of CPU scheduling algorithms: deterministic modeling, queuing analysis, and simulators. Deterministic modeling uses a predetermined workload to directly compare algorithms, but only applies to that specific scenario. Queuing analysis uses formulas like Little's Law to calculate metrics like wait time given arrival and service rates. Simulators provide the most accurate evaluation by programming a computer system model and gathering statistics from executing different algorithms on sample processes and workloads.

Uploaded by

Indian Best
Copyright
© © All Rights Reserved
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 2

Unit II

Performance Comparison
The performance of the CPU scheduling algorithms depends on the variety of features
including the probability distribution of service times of various process, the efficiency of
the scheduling and content switching mechanisms, and nature of the I/O demand and the
performance of the I/O subsystem.

Deterministic modeling
This method takes a particular predetermined work load and defines the performance of
each algorithm for that work load. For example the work load is given as: All 5 processes
arrive at time 0. The length of CPU burst time is given in milliseconds.

Process CPU Burst Time


(in milliseconds)
P1 6
P2 12
P3 1
P4 3
P5 4

Consider FCFS, SJF and RR (Quantum = 1) scheduling algorithms for this set of
processes. The algorithm which would give the minimum average time, will be the best
one. This type of algorithm execution is said to be deterministic modeling. It is simple
and fast. It requires exact numbers of input and its answers apply to only those cases.

Queuing analysis
Queuing analysis can be useful in comparing algorithms. If we know the arrival rates and
service rates, we can compute CPU utilization, average queue length, average waiting
time and so on. This area of study is called queuing analysis. For example consider the
Little’s formula
N=⋋ XW
Where, N = Average queue length
⋋ = Average arrival rate in the queue
W = Waiting time of the process
We can use Little’s formula to compute one of the three variables, if we know the other
two.

Simulators
We can get more accurate evaluation of scheduling algorithms with simulators.
Simulation involves programming model of the computer system. Software and data
structures represent the major components of the system. When the simulation executes,
statistics that indicates algorithm performance are gathered. For example consider the
following figure:

Actual Comparison
16 Performance 1
process
execution
Data
Performance
Simulation statistics for FCFS
for FCFS

Performance statistics
Simulation for SJF
for SJF

Performance statistics
Simulation for RR
for RR

Fig: Evaluation of CPU scheduling by simulation

16 Performance Comparison 2

You might also like