ECE 366: Computer System Performance Evaluation

Spring 2001

Lecture: M-W-F 12:00 – 12:50, Gr II, Room 222

Lab: T-R 12:30 – 1:45, Gr II, Room 222

Syllabus

 

Instructor

Dr. Howard Michel

Office: Gr II, Room 221A

Phone: 508-910-6465

Email: hmichel@umassd.edu

Web Page: http://www.ece.umassd.edu/Faculty/hmichel/home.htm

Office Hours: M – W 11:00 – 12:00 PM, T-Th, 10:00 – 11:00 AM

The above office hours are the minimum that we will be available. Feel free to stop in any time my door is open, email me questions, or call for an appointment.

Course Overview

This course will take a rigorous look at the various components—both hardware and software—that comprise a computer system. The emphasis of this course will be on the interactions among these components and the external environment, and how these relationships effect the performance of the computer system. This approach will separate computer science professionals, electronic engineers and computer hobbyist from computer engineers. The pace of this course will be suitable for junior or senior computer engineering students—that is, expect to spend 3-4 hours per week outside class for every hour in class.

 

Prerequisites

MTH 331 Probability

ECE 262 Computer Design

ECE 161 or CIS215 Computer Programming with C

It is expected that you have developed a technical maturity that includes a basic understanding of general computer organization concepts and are comfortable programming in the C programming language. We will not cover these materials again in this course, but rather build upon these skills. If you are in doubt about your preparation for this course, please ask the instructor.

Course Objectives

My primary objectives for the students in this class are:

  1. To understand the various components that comprise a computer system and their effect on total system performance.
  1. To demonstrate an ability to apply the correct tools to computer system performance problems.

Materials

Text:

  • Capacity Planning and Performance Modeling: From Mainframe to Client-Server Systems, by Daniel A. Menasce, Virgilio A. F. Almeida, and Larry W. Dowdy, Prentice Hall, 1994
    •

    References:

    1. The Art of Computer Systems performance Analysis, by Raj Jain, Wiley, 1991
    2. Introduction to Computer Syatem Peformance Evaluation, by K. Kant, McGraw-Hill, 1992
    3. Simulation, Modeling and Analysis, 3rd edition, by Averill M. Law and W. Davis Kelton, McGraw-Hill, 2000
    4. Probability, Statistics, and Queueing Theory with Computer Science Applications, 2nd edition, by Arnold O. Allen, Academic Press, 1990

     

    Grading Policies

    1. Grades will be computed on the standard A+ to F scale. The weights assigned to the various items are:

    2. Assignments are due at the beginning of the specified class. Late assignments will be penalized at 10% per 24-hour or fraction thereof. No assignment will be accepted for credit after it has been returned to the class.

    3. All work to be handed in must be done individually unless the instructor explicitly states otherwise. Any assignment or portion of an assignment that is duplicate or group result will receive a grade of zero for all parties concerned.

    4. Makeup exams will only be given in the case of serious illness or emergency. You must personally contact the instructor before the scheduled exam.

    Academic Honesty

    All work turned in is expected to be your own work. Do not use another student’s work or give your work to others. Do not leave your work lying around where other students can copy it. Any assignment or program that is duplicate or team effort will receive a grade of zero for all students concerned regardless of the reason.

    You may discuss concepts or material covered in class, but do not discuss solutions to assigned problems. Specifically, you are not allowed to share any details of your solution, including algorithms or program code. You may help someone else debug their program as long as you do not start substituting in your own code or algorithm. A better alternative is to jointly solve another problem.

     

     

    ECE 366: Computer System Performance Evaluation

    Spring 2001

    Lecture: M-W-F 12:00 – 12:50, Gr II, Room 222

    Lab: T-R 12:30 – 1:45, Gr II, Room 222

    Topic Schedule

    The proposed topic schedule is tentative. Topics may actually be covered earlier or later than indicated depending on the abilities and requirements of the class. Please keep up with the reading as determined by the material covered in class.

    Test dates are not expected to change. Test content will be adjusted to reflect material actually covered.

    Week

    Topic

    Readings

    1

    		 Introduction, review Ch. 1

    2

    		 Capacity planning methodology Ch. 2.1 – 2.6

    3

    		 Workload forecasting, performance prediction Ch. 2.7 – 2.9

    4

    		 Performance models: basics, interactive systems, and batch systems Ch. 3.1 – 3.4

    5

    		 Performance models: multi-class systems

    Exam 1: March 2, 12:00 – 12:50

    		 Ch. 3.5 – 3.10

    6

    		 System level models Ch. 4.1 – 4.3

    7

    		 System level models Ch. 4.4 – 4.7

    8

    		 Solution techniques: basics Ch. 5.1 – 5.3

    9

    		 Solution techniques: convolution, mean value analysis, other methods Ch. 5.4 – 5.8

    10

    		 Multiple-class models: 2-class model, notations, closed models

    Exam 2: April 10, 12:00 – 12:50

    		 Ch. 6.1 – 6.4

    11

    		 Multiple-class models: open models and mixed models Ch. 6.5 – 6.7

    12

    		 Performance of client-server architectures: intro, workload characterization, Ch. 7.1 – 7.5

    13

    		 Performance of client-server architectures: different performance models Ch. 7.6 – 7.10

    14

    		 Practical computer systems: memory and disks Ch. 8.1 – 8.3

    15

    		 Practical computer systems: CPU, multiprocessors Ch. 8.4 – 8.7

     

    Final Exam:

    Thursday, May 24, 8:00 AM – 11:00 AM or

    Wednesday, May 23, 11:30 AM – 2:30 PM (class can decide)