return to Kris' Home Page
An Ad Hoc Approach to
Undergraduate
Curriculum Development in
Computational
Science
DOE High Performance Computing
Education Conference
Albuquerque NM
February 10, 1994 (updated)
Kris Stewart
Associate Professor
Department of Mathematical Sciences
San Diego State University
San Diego CA 92182-0314
Computational Science Curriculum Coordinator
San Diego Supercomputer Center
A National Laboratory for
Computational Science and Engineering
stewart@cs.sdsu.edu
At a state university, such as San Diego State
University (SDSU), it is difficult to develop large,
comprehensive academic programs. With the last few
years of severe budget crisis in California, this has
become even more difficult. Still, there is a need to
incorporate changing technology and prepare students to
satisfy the needs of local industry. One major
advantage for SDSU is the fact that it was one of the
original consortium members of academic institutions
when the National Science Foundation founded the five
centers for high performance computing. Having one of
the national centers in your own town is an advantage
for any state school (though easy network access is
changing this).
Two courses have been accepted into the SDSU
undergraduate catalog recently and I wish to describe
them in this paper. Additionally, I will highlight
other nationwide programs providing curriculum
materials, freely, over the Internet. The first course
will be described in rough terms only because it has
not yet been taught. Due to the backing of the San
Diego Supercomputer Center, CS 205 Introduction to
Computation Analysis will be taught Fall Semester 1994.
======================================================
CS 205 Introduction to Computational Analysis
The intent of this sophomore level course is to
reinforce calculus and begin a student's introduction
to programming and floating point arithmetic in a fo-
cussed manner. The focus is on the skills that are
essential to students in the sciences if they plan to
pursue a degree in Computational Science.
Many students become familiar with word proces-
sors, spreadsheets and communications packages in high
school or at home due to the wide spread use of per-
sonal computers. The fundamental skill missing from
the use of these standard software tools is an under-
standing of floating point arithmetic. This includes
estimating errors, estimating rates of convergence,
programming in a manner to avoid floating point excep-
tions like overflow/underflow and much more.
Familiarity with floating point arithmetic is
essential in computational science whether students
intend to do their own programming or they intend to
use discipline-specific computational packages. The
proposed course would illustrate this arithmetic using
numerical approximations derived directly from Taylor
Series, thereby reinforcing the calculus training that
students have just received in their freshman year.
The programming environment will be MATLAB. This
is a powerful software tool that facilitates learning
Linear Algebra concepts, but MATLAB would be used here
for its other properties.
o available and uniform over different computing
platforms, including a very affordable student edition
which students with home computers could purchase
o graphics capabilities are reasonably simple to use
and easily extend to publication quality graphics
o programmability: although MATLAB has an extensive
interactive capability to compute solutions for most
standard problems of linear algebra in a numerically
stable manner, MATLAB also contains a subset of the C
programming language that can be used to introduce
students to programming.
More details on this course will evolve as Fall
Semester gets closer. Just as for the CS 575 course
described below, the course materials will be made
available via anonymous ftp once they exist.
======================================================
National Programs Providing Curricula Materials over
the Internet
Before describing the supercomputing course, which
is a well developed course, I would like to highlight
some alternate programs that I have either participated
in, or have just been told of. My association with the
San Diego Supercomputer Center has been very helpful in
this regard, since SDSC tends to attract computational
scientists and the SDSC staff will then tell me of
other work going on.
A recent example is Dr. Ian Johnston at the School
of Physics at the University of Sydney, NSW 2006,
Australia. We have just begun communicating
electronically and he has mailed (postal type) some
articles of his concerning his work with M.U.P.P.E.T
(Maryland University Project for Physics Education
Technology).
I was also fortunate to attend the faculty
workshop given by Dr. Lloyd Fosdick and Dr. Liz Jessup
in the summer of 1992 at the University of Colorado,
Boulder. Dr. Jessup is also on this panel and will
provide more details. The course notes are extensive
and well developed and are available via anonymous ftp
from ftp.cs.colorado.edu in the directory pub/HPSC, I'd
recommend retrieving the README as well as the UNIX
compressed, postscript file, SciCompOverview.ps.Z.
In the last year, I have joined a DOE funded
program from Ames Laboratory call UCES (Undergraduate
Computational Science and Engineering). Quoting from
the program's flyer: "This project strives to promote
the emerging field of computational science as an
interdisciplinary/multidisciplinary approach to
scientific analysis. Currently our emphasis is on
collecting, developing and distributing a set of
educational materials in computational science." For
more information on the program, contact Dr. Tom
Marchioro (uces@uces.ameslab.gov). The first course to
emerge from this work is a freshman level course on
Engineering Computing at the University of Utah,
developed and taught by Dr. Chris Johnson and Dr. Joe
Zachary (who are at this conference). The materials
are available via anonymous ftp from cs.utah.edu and
form a tutor system which runs in conjunction with
Maple to introduce students to the GUI environment;
UNIX; computing, programming and visualization using
Maple; as well as C and Fortran programming. This
program is listed in the October 1993 issue of Computer
with its cover story on "Computational Science &
Engineering".
Another program which I heard of from Dr. Robert
Borrelli is the Consortium for Ordinary Differenential
Equations Experiments (C*ODE*E) from Harvey Mudd
College. This group produces a newsletter (email
codee@hmc.edu to subscribe to the hardcopy or use
anonymous ftp to ftp.hmc.edu in the directory
/pub/CODEE to obtain the postscript files of back
issues.)
One of the recent issues described a program
through St. Olaf College on modeling. There are
currently eight complete projects available via
anonymous ftp to ftp.stolaf.edu which cover:
Eulerian Wobble
First Order Bifurcations
Pursuit Differential Equations
Building a Model: Predator-Prey Simulations
Competing Species with Harvesting
Regular Perturbation
Regular Perturbation II: The Precession
of Perihelion of Mercury
The Salt in the Tank Problem
Each postscript file contains a problem description,
notes to instructor, solution code in Maple and figures
of the computed solutions. These were prepared by Dr.
Steven Kennedy at St. Olaf and his group.
As a last item in my ramblings here, I present my
current favorite NCSA Mosaic sites:
http://gopher.sdsc.edu/Home.html SDSC's Home Page
http://gopher.sdsc.edu/SR93/SDSC_SR.html 1993 Science
Report from SDSC
http://compsci.cas.vanderbilt.edu/csep.html Computa-
tional Science Education Project (CSEP) the graduate
computational science program supported by DOE.
http://www.w3.org/hypertext/DataSources/Yanoff.html
Yanoff's List
http://www.gatech.edu/gvu/gvutop.html Introductory
Graphics and Visualization
http://galaxy.einet.net/GJ/inidex.html Traditional
structure for Galaxy's Gopher Jewels
http://galaxy.einet.net/gopher/gopher.html Searchable
form for Galaxy's Jewels
file://dggis.rtpnc.epa.epa_ceam/wwwhtml/ceam_home.html
EPA Center for Exposure Assessment Modeling
http://www.exploratorium.edu/ SanFranciscoExploratorium
========================================================
CS 575 Supercomputing for the Sciences & Engineering
The second course at SDSU, is CS 575 Supercom-
puting for the Sciences. This course is being taught
this semester for the third time, the first two times
as an "experimental CS 596" course, which can be
offered at a department's discretion. The course
appeared in the 1993/94 Undergraduate catalog after the
campus wide scrutiny process was satisfied. This
course was supported by a grant from the Divsion of
Advanced Scientific Computing at NSF. It was coupled
with a workshop offerred in the summer for faculty,
from primarily undergraduate institutions, who were
interested in incorporating High Performance Computing
and Communications into their undergraduate curricula.
These workshops, called SUE (Supercomputer and
Undergraduate Education) were offerred in summer 1991
and 1992. These faculty workshops also received
supplementary funding and support from the Cray
Research Foundation.
The course CS 575 focusses on the Cray vector
supercomputer and uses the text, Computer Architec-
ture: A Quantitative Approach by Hennessy and
Patterson. It is an interdisciplinary course to
introduce students in the sciences and engineering to
advanced computing techniques using the Cray C90
Supercomputer at the San Diego Supercomputer Center.
Prerequisites are an extensive programming background
in Fortran or C.
Students are graded based on computing projects
(75%) and one take home midterm (25%) based on material
from the text by Hennessy and Patterson.
The actual syllabus of the course being taught Spring
1994 is given below:
[UNIX] The Cray is a UNIX machine. We begin the
semester using the SUN SPARC 10 machine while students
become reasonably functional with UNIX and electronic
mail. Students are given a demonstration of using
gopher to access copies of the lectures notes which are
available from ftp.sdsc.edu (1 week)
There is a salmon-colored handout available at this
meeting describing how to access all the lecture notes
via anonymous ftp. This is also included as an
Appendix to this paper.
[Architecture] To effectively use the Cray, you need an
understanding of the hardware features of the machine
that make it so fast. We use the text by Hennessy and
Patterson to gain an INTUITIVE understanding of the
architecture of a vector supercomputer. (6 weeks)
[Chapter 1] Fundamentals of Computer Design
[Chapter 2] Performance and Cost
[Sect. 3.3] Operand Storage in Memory -
Classifying General-purpose Register Machines
[Sect. 3.7] The Role of High-Level Languages and
Compilers
[Sect. 4.5] The DLX Architecture
[Chapter 6] Pipelining
Goal->[Chapter 7] Vector Processors
[Sect. 8.1] Memory-Hierarchy Design; Introduction
[Sect. 8.2] General Principles of Memory
Hierarchy
[Sect. 8.4] Main Memory (Note - Cray is not a
virtual memory machine)
[Sect. 8.7] More Optimizations Based on Program
Behavior
[Computers and Ethics] Guest lecturer from SDSC visits
the class. There is an essay assignment (which
is graded on a pass/fail basis) (1 week)
[Performance Evaluation/Usage Monitoring] The class
starts working on the Cray at this point.
Students are frequently reminded that they have a
FIXED amount of Cray time to use to complete their
computer projects. Student must learn to manage
their allocation of a valuable, but scarce,
resource. (2 weeks)
[Impact of Architecture on Algorithms] Cray-specific.
(2 weeks)
[Software Tools] Effective use of the Fortran or C
compilers. Debugging and performance tools. (3
weeks)
========================================================
Major Themes of the SUE Program (faculty workshop and
course at SDSU)
1 Interdependence of Computer Science & Scientific
Experts
The workshop faculty represented two groups: those
interested in learning discipline-specific applications
packages and those interested in using software tools
to facilitate programming. To accommodate both groups,
we presented an overview of available resources
(applications packages and program optimization
techniques) and encouraged faculty to seek further
information independently. We identified the technical
people they could contact for further information on
their particular interests.
Similarly, the required programming background for the
course reinforced the traditional separation between
computer science and science/engineering students by
encouraging the former to take the course. In both
cases, the participants interacted well as they
realized the benefits of working with others with
different strengths.
2 CRAY architecture
The course uses the text Computer Architecture: A
Quantitative Approach, by Hennessy and Patterson
(published by Morgan Kaufmann) because the model
developed for the Cray 1 provided a great deal of
insight into the design of the Cray C90.
We try to gain an appreciation for the sources of the
Cray C90's power and understand the subtleties of its
design as they impact the way a programmer should
approach a problem. We carefully avoid many of the
sections of the text that do not directly relate to the
Cray C90. The text is supplemented with documents by
Cray Research, Inc.; particularly the following titles:
TR-OPT (Rev. D) cf77 & scc Features and Optimization.
An excellent, Cray-specific training report covering
the Fortran (cf77) and C (scc) compilers. Provides
code examples, diagrams, and explanations of crucial
vectorization topics (and the conditions that inhibit
the compiler), memory organization, performance tools,
common optimization techniques, and much more.
TR-YSAAP Cray Y-MP System Architecture for Applications
Programmers. Covers material at the assembler level in
more detail. Works well with the models developed in
Hennessy and Patterson (unfortunately this document is
no longer available)
3 Communications
Both the workshop and the course covered the resources
available through the Internet, including:
a. Accessing sources of information
(news groups, anonymous FTP):
oak.oakland.edu (IBM PC software)
sumex-aim.stanford.edu (Macintosh softw)
ftp.cs.titech.ac.jp
(Dr. David Kahaner's Japan Bulletins)
b. Communicating with peers (via e-mail)
c. Using FTP to transfer programs between
machines using gopher.
4. Measuring Performance
At many universities, accounts issued to students in
computer courses are limited only by the amount of
available file space. Therefore, the concept of
monitoring CPU usage can be new. The students in the
SDSU class are given programming projects on a
mainframe at SDSU and are introduced to the concept of
accounting using crude UNIX timing tools (e.g., dtime)
before they move to the Cray.
When students program on the Cray, they have a fixed
amount of CPU time to work with. Students are reminded
that they must complete their course projects without
exceeding their CPU allocations. The more sophisticated
timing and resource monitoring tools are used to ensure
that the student projects do not use up more time than
they are allocated for the semester.
5 Computer ethics/responsibility
Students in the course choose between two computer-
related scenarios and write a one-page essay on a
scenario, discussing the behaviors of the individuals
involved. The essays were graded on a pass/fail basis
(only essays that showed a lack of thought or effort
"failed"). The goal, for the instructor, was to gain a
better understanding of the students' attitudes
concerning computer-related issues.
Dr. Dan Sulzbach, Executive Director of SDSC, has
given, in the past, a week of lectures on computer
ethics, which stimulated some very interesting class
discussions.
Dr. Sulzbach began his talk by stating:
"I am not a philosopher, not an ethicist, not an
expert. I am a computer professional like you. I'm
not here to preach because I have no license or
authority to do that. I hope only to raise some
issues related to computer ethics. I will undoubtedly
ask more questions than I answer."
Computer accounts on the Cray C90 are distributed only
after the week-long discussions on ethics.
6 Writing
Most of the students have no problem with the computer
ethics essay, but they typically have had little
experience in writing up a science-oriented programming
report. Therefore, the assigned programming project
consists of the following components:
a. Solve a problem on a campus mainframe and document
the mainframe's performance.
b. Solve the same problem on the Cray.
c. Compute an enhancement of the problem on the Cray
and document the Cray's performance.
Instructor feedback after reading the document from
assignment (a) greatly enhanced the organization and
content of the final document in (c).
Main Project
Science-Oriented Program to be run on SDSU Mainframe
and subsequently on the CRAY
The main project is crucial to this course. Students
will use the Cray to run this project, and the
instructor does not want them to squander Cray
resources before they became familiar with their
problems. Students are allowed to pick from a selection
of problems provided by the instructor, or they could
solve a scientific Cproblem from their particular
backgrounds or work environments.
Good sources for problem statements were:
"Computing Applications to Differential Equations:
Modelling in the Physical and Social Sciences," by
J.M.A. Danby; Reston Publishing
"Numerical Methods and Software" by Kahaner, Moler
and Nash; Prentice-Hall Publishers.
"Computational Physics" by Koonin and Meredith;
Addison Wesley Publishing Company.
The MAIN PROJECT assignment:
a) Get your "science" program running on the SDSU
mainframe.
b) Write a report describing your problem and your
program's performance on the SDSU mainframe.
c) Get your program running on the Cray.
d) Extend your problem. For example, use a finer grid
spacing or use more species in an interaction. (This
will depend on your particular problem.)
e) Submit a final report on your Cray project.
Topics to be covered in your report:
a) Your write-up should have a self-contained statement
of the problem. The reader should not have to read your
code to find out what equations you are working with,
or what the specific problem is that you are solving.
b) Give a complete reference to where the problem came
from.
c) Define your measure of work so that comparisons can
be made when you run on the Cray. You can't talk about
"faster" or "better" without a specific measure of
performance.
d) Discuss what conclusions are drawn from the problem
itself. What is the "science" story revealed by the
original problem? Why was this problem solved.
e) You should carefully organize the results. A summary
of pertinent results for both the "science" of the
problem and the "performance" of the program should be
presented. Optionally, include an appendix for more
detailed results.
========================================================
Appendix: Anonymous FTP Access to Course Notes
Feb. 3, 1994
Kris Stewart, Assoc. Prof., Dept. Math. Sciences, SDSU
Computational Science Curriculum
Coordinator, SDSC
SUE = Supercomputing and Undergraduate Education
SDSU = San Diego State University
SDSC = San Diego Supercomputer Center
As described in the article "SUE workshop" in Vol.
8, No. 11-12 (Nov.-Dec. 1992) of Gather/Scatter from
SDSC, the lecture notes from the SUE workshop offerred
last summer at SDSC are available to the public. You
can access them through the Internet by following the
following directions
ftp ftp.sdsc.edu (establish ftp contact with SDSC
machine)
You will be prompted for the standard ftp logon. You
need to identify yourself as "anonymous" and enter your
complete email address when prompted for a password.
Name: anonymous
Password: type your complete email address here
NOTE - every directory level has a README file
summarizing the contents in that directory
cd pub (change directory)
cd sdsc (change directory)
cd docs_and_education (change directory)
cd education (change directory)
get README (you can later examine this file
for more details on the organization of this info)
get super92.asc (this is an ascii copy of Stewart's
Poster Session on SUE at Supercomputing '92)
You may want to just stop here. There is a lot
more information available. I feel the files in the
directory "SUE-notes" are in adequately good shape. If
you would like to continue to retrieve more documents
cd SUE-notes (change directory)
get README (this file gives an overview)
Again, you may want to stop here and look at the
somewhat lengthy file README in the SUE-notes
directory. If you want to get it all:
prompt (turn off the automatic prompting)
mget * (get all the files in the directory SUE-notes)
bye (terminate ftp session)
This will copy a large amount of information to your
own account on your home computer. You may edit the
files at well, but you should give credit to their
original source.
To make things simpler, I'm including the README
file mentioned above in this mailing. -kris
------------------------------------------------------
README in pub/sdsc/docs_and_education/education
available via anonymous ftp to ftp.sdsc.edu
NAME
supercomputers_in_education Information on
using SDSC computers for courses
super92.asc ASCII document of Dr. Kris
Stewart's handout in conjunction with her poster
session at Supercomputing '92 in Minneapolis. This
gives an overview of the SUE activities at SDSC as well
as descriptions of educationally related files
available in the following directories:
STEP-notes A directory of files from the
Supercomputer Teacher Enhancement Program (STEP).
SUE-notes A directory of files from Dr.
Kris Stewart's SUE faculty workshop offerred 1991 and
1992 at SDSC - THESE ARE MOST UP TO DATE - USE FIRST.
SUE stands for Supercomputing and Undergraduate
Education.
Supercomputing-Course-Notes A directory of
lecture notes from Dr. Kris Stewart's course CS 575
Supercomputing for the Sciences taught at SDSU in 1991
and 1992. This is an interdisciplinary course for
students with a good programming background in Fortran
or C focussed primarily on using the Cray Y-MP 8/624 at
SDSC. CAVEAT - the files in this directory are not in
as good of shape as those in the directory SUE-notes.
I recommand the files in the SUE-notes directory more
than those in this directory. Examine these with
caution.
DESCRIPTION pub/sdsc/docs_and/education/education
is a subdirectory of SDSC's anonymous ftp archive. It
contains education materials.
CONTACT SDSC Consultants (consult@sdsc.edu) (619)
534-5100
ADDITIONAL INFORMATION For additional information
about SDSC's anonymous ftp, see the files in the help
directory.