YCP Logo Syllabus

Instructor:David Hovemeyer
Office:KEC 113
Phone:717-815-6582
Office Hours:MWF 10-11 AM, T/Th 2-3 PM
Email:dhovemey@ycp.edu
Class Location:KEC 124
Class Times:T/Th 12:30 - 1:45 PM

CS 497 - Parallel and Distributed Computing

Spring 2009

Updated Jan 27th - Schedule revised

Updated Jan 29th - Schedule revised

Updated Feb 10th - Added reading for Weeks 4/5

Course Description

This course will explore the theory and practice of designing and implementing programs to run on multiple processors in parallel. A number of programming projects using C with the pthreads and MPI libraries will provide an opportunity to implement and benchmark parallel programs.

Prerequisites

CS 201. At least one upper level (300 or above) computer science course with a significant programming component is strongly recommended.

Course Structure and Expectations

This main goal for this course is to introduce the theory and practice of designing and implementing parallel programs. A parallel program executes on multiple processors or computers simultaneously. For problems that are amenable to parallelization, a parallel program can complete a computation much faster than a sequential program.

This course will have challenging programming assignments. They will be challenging for several reasons:

  • Designing parallel algorithms is hard. We will cover a number of good strategies for designing parallel algorithms, but you should expect to spend significant time up-front using pencil and paper to design an algorithm for your program before you start coding.
  • Reasoning about the behavior of parallel programs is much more complicated than reasoning about the behavior of sequential programs. You should expect to spend significant time and energy debugging your programs.
  • Threading and message-passing libraries, such as pthreads and MPI, are complex. You should expect to spend a significant amount of time reading the text and online resources to get up to speed.

For all of these reasons, it is absolutely essential that you start programming assignments early. I will give you plenty of time to complete each assignment, but if you wait until the last minute to start, I will not make a heroic effort to help you complete it. Generally, I will ask you to submit a written design, with an outline of your algorithm, before you start the implementation.

We will be writing programs in the C programming language. If you do not know C well, you will need to get up to speed quickly. I recommend any of the following books:

  • Samuel P. Harbison, Guy L. Steele, C: A Reference Manual. If you're going to buy one C book, this is an excellent one.
  • Peter Prinz, Tony Crawford, C in a Nutshell. Another very good general reference on C.
  • P.J. Plaugher and Jim Brodie, Standard C: A Reference. An excellent quick reference. Any edition is fine.
  • Peter Prinz, Ulla Kirch-Prinz, C Pocket Reference. Another good quick reference.

We will be developing and running programs on computers running Linux. You may wish to buy a general introduction to Linux or Unix.

Learning Outcomes

By the end of this course, you will be able to:

  • Design and implement parallel programs using threading and message-passing methods of communication between parallel tasks
  • Run parallel programs on multiprocessors and clusters
  • Understand and analyze the performance of parallel algorithms and programs

Policies

Grades

Grades are assigned on a 100-point scale:

Numeric Range Letter Grade
90-100 A (4.0)
85-90 B+ (3.5)
80-85 B (3.0)
75-80 C+ (2.5)
70-75 C (2.0)
60-70 D (1.0)
0-60 F (0.0)

Your overall grade for the course will be determined as follows:

  • Programming assignments: 40%
  • Two midterm exams: 30% (15% each)
  • Final exam: 15%
  • Written homework and short programming exercises: 5%
  • Quizzes: 5%
  • Attendance and participation: 5%

Course website

Please check the course web page, http://faculty.ycp.edu/~dhovemey/spring2009/cs497/, regularly for important announcements.

Reading Assignments

Reading assignments will be posted on the course web page. I expect you to do the reading before class. When I give a lecture, I will assume you have done the reading. I encourage you to use class time to ask questions about parts of the reading you did not understand to your satisfaction.

Posting and submission of assignments and labs

Assignments will be posted on the course web page, http://faculty.ycp.edu/~dhovemey/spring2009/cs497/.

Assignments will be submitted using the server https://camel.ycp.edu:8443. You will receive an email containing the username and password you should use for this server.

Academic Integrity

The college catalog states the following:

Academic dishonesty will not be tolerated at York College. Academic dishonesty refers to actions such as, but not limited to, cheating, plagiarism, fabrication of research, falsification of academic documents, etc., and includes all situations where students make use of the work of others and claim such work as their own.

Please refer to the college catalog for an explanation of the official college policies relating to academic integrity.

The following policy pertains to homework and graded (individual) programming assignments in this course:

All homework assignments and graded (individual) programming assignments are to be completed individually. I encourage you to discuss high level concepts and strategies with other students, but any work you submit must be yours alone.

Direct copying of code or other work from other students, web sites, or other sources is absolutely forbidden under any circumstances.

Any sources (books, websites, articles, fellow students, etc.) that you consult in completing an assignment must be properly acknowledged. In general, I strongly discourage you from using any resource not explicitly listed in the course syllabus or on the course web page. When you work on a programming assignment, it must be your program, not your adaptation of someone else's program.

In order to receive credit for an individual homework assignment or programming assignment, you must submit a signed attestation confirming that you completed the assignment on your own, and did not copy any code or other material from any person or source.

Quizzes and exams must be completed individually.

Any violation of the course's academic integrity policy will be referred to the Dean of Academic Affairs, and could have consequences ranging from a 0 on an assignment to dismissal from the college.

Late Assignments

Late assignments will be marked down 5% per day late. No credit is given for assignments turned in after grading has taken place.

Exams

No make-up exams will be given without approval of the instructor prior to class unless proof of extreme emergency or illness is provided. All exams will be open book and open notes.

Attendance and Participation

I expect you to attend class and participate regularly in class activities. If you miss a class, please notify me in advance. You are responsible for all material covered in class, regardless of whether or not you were present. If you attend and participate in class regularly, you can expect to receive full credit for attendance and participation. Frequent absence and/or lack of participation in class activities will reduce the credit you receive for attendance and participation.

Use of Personal Technology in the Classroom

While York College recognizes students’ need for educational and emergency-related technological devices such as laptops, PDA’s, cellular phones, etc., using them unethically or recreationally during class time is never appropriate. The college recognizes and supports faculty members’ authority to regulate in their classrooms student use of all electronic devices.

Communication Standards

York College recognizes the importance of effective communication in all disciplines and careers. Therefore, students are expected to competently analyze, synthesize, organize, and articulate course material in papers, examinations and presentations. In addition, students should know and use communication skills current to their field of study, recognize the need for revision as part of their writing process, and employ standard conventions of English usage in both writing and speaking. Students may be asked to further revise assignments that do not demonstrate effective use of these communication skills.

Schedule

This schedule may be subject to minor changes. Any changes will be announced in class and also be posted on the course web page.

Key to reading assignments:

  • Pacheco is the assigned textbook
  • PTP is the POSIX Threads Programming tutorial at Lawrence Livermore National Laboratories.

Updated Jan 27th: Switched weeks 2 and 3

Updated Jan 29th: Moved first Pacheco reading to week 6

Updated Feb 10th: Added reading for Week 4/5.

Week Note Topic Reading
Week 1: Jan 19-23 No class Tuesday Course overview Pacheco, Chapter 1
Week 2: Jan 26-30   C review; threads, pthreads PTP, Sections 1-5
Week 3: Feb 2-6   Thread synchronization;
Concurrent data structures
PTP, Sections 6-7
Week 4: Feb 9-13   MT design patterns;
Races and deadlocks
Doug Lea, A Java Fork/Join Framework
Week 5: Feb 16-20   Threads wrap-up  
Week 6: Feb 23-27   Overview of Parallel Computing Pacheco, Chapter 2
Week 7: Mar 2-6 Winter break: no class
Week 8: Mar 9-13   Intro to message passing, MPI Pacheco, Chapter 3
Week 9: Mar 16-20   Simple MPI Applications Pacheco, Chapter 4
Week 10: Mar 23-27   Collective Communication Pacheco, Chapter 5
Week 11: Mar 30-Apr 3   Grouping Data Pacheco, Chapter 6
Week 12: Apr 6-10 No class Thursday Communicators, Topologies Pacheco, Chapter 7
Apr 9-13 Spring break: no class
Week 13: Apr 13-17   Performance Pacheco, Chapters 11-12
Week 14: Apr 20-24   Distributed Computing?
TBA
Week 15: Apr 27-May 1      
Week 16: May 4-8