Computer Systems Performance Measurement and Evaluation
Course Objectives:
- Learn to use appropriate statistical techniques to compare systems and interpret measured data.
- Learn how to develop and apply measurement tools and techniques.
- Learn how to use analytical modeling.
- Learn how to appropriately design experiments.
- Learn how to develop and use various types of simulations.
- Learn to choose an appropriate performance evaluation technique.
Instructor:
David J. Lilja
- Office: 6-131 Keller Hall.
- Email: lilja at umn dot edu
- Office hours: Click here to schedule a meeting with me.
Required Texts:
- "Measuring Computer Performance: A Practitioner's Guide," David J. Lilja, Cambridge University Press, 2000, ISBN 0-521-64670-7 (paperback).
The textbook is available at the bookstore in Coffman Union. An elecronic copy also is available from the Library Reserves through the class moodle site.
Errata -- This is the list of corrections for the book. Please let me know if you find any other errors. - "Computer Architecture Performance Evaluation Methods," Lieven Eeckhout, Morgan and Claypool Publishers, 2010, ISBN 978-1-60845-467-9.
- Papers from the supplemental reading list shown in the syllabus.
As a student registered in this course, you can access these papers through the course moodle site.
Course Details:
- Look at this web site for important University policies that affect your participation in this class.
Homework Assignments:
- The homework assignments will be posted on the class moodle site as they are assigned.
Resources for improving your writing, talks, and poster presentations:
- University of Minnesota Center for Writing
- "How to Give an Academic Talk, v3.1," by Paul N. Edwards (University of Michigan)
- Poster Perfect from The Scientist magazine
- Creating an Effective Scientific Poster Presentation from the University of Minnesota
- Creating posters in PowerPoint from the University Libraries.
- Effective Poster Design from the University of Guelph
Other resources:
- A program execution timer that can be adapted for most Linux machines - unixtimer.c
- Open source statistical software - R Project
- Linear Regression Using R, by David J. Lilja, available for free download from the University of Minnesota Libraries.
- Introductory Statistics with R, by Peter Dalgaard, available through the library to registered students.
- A nice (free) text on design of experiments - A First Course in Design and Analysis of Experiments
- A database of processor characteristics and performance results - CPU DB