CS4481 – System Performance Analysis
Spring 2007

Objectives: As a key step in selection of processors, operating systems, database, system networks, and languages, performance analysis plays an integral role in the design and procurement of new computer systems. This course introduces some fundamental performance analysis techniques including methods for performance measurement, monitoring, experimental design, system simulation and modeling. The students will learn in this course:

1.       comparing two or more systems

2.       system tuning

3.       performance bottleneck identification

4.       characterizing the load on the system (workload characterization)

5.       determining the number and size of components (capacity planning)

6.       predicting the performance at future loads (forecasting)

7.       system simulation and monitoring

8.       queuing theory and modeling

 

Text: Raj Jain, The Art of computer Systems Performance Analysis: Techniques for experimental Design, Measurement, Simulation, and Modeling, Wiley, 1991.

Instructor: Zhenlin Wang

Office hours: Rekhi 207, MW 3-4:30 PM or by appointment

E-mail: zlwang AT mtu.edu


Grading:

Homework

20%

Projects

60%

Midterm (in class)

15%

Class Participation

5%

 

Topics:

Measurement techniques and tools, workload selection and benchmarking

Experimental design and analysis

Simulation, random number generation

Queuing networks