Parallel Computing for Partial Differential Equations - Syllabus
Universität Kassel - Sommersemester 2019
Matthias K. Gobbert - University of Maryland, Baltimore County
This page can be reached via my homepage at
http://www.umbc.edu/~gobbert.
Basic Information
- Matthias K. Gobbert, University of Maryland, Baltimore County,
Office AVZ Room 2419, (05 61) 8 04-46 77, gobbert@umbc.edu
- The Vorlesungsverzeichnis of Sommersemester 2019 lists the
course as BV21 and MV40 with 2+1 SWS.
The course will be taught in block form
during the two weeks from June 17 to July 03, 2019.
The first class meeting is Monday, June 17, 2019, at 09:15 a.m. at AVZ.
If you are interested in the course, please register in its
Moodle course,
so we can contact you.
See the detailed schedule
for more information on the contents of the course.
- Prerequisites: Numerik I,
recommended preparation:
proficiency in programming C/C++ and
in using the Unix/Linux operating system;
or consent of instructor
- Books on parallel computing, the programming language C, and Matlab:
-
Required textbook on parallel computing:
Peter S. Pacheco,
Parallel Programming with MPI,
Morgan Kaufmann, 1997.
Associated webpage:
http://www.cs.usfca.edu/~peter/ppmpi.
We will have explicit reading assignments for several chapters
from this book at the beginning of the semester.
-
Recommended book on the programming language C:
Brian W. Kernighan and Dennis M. Ritchie,
The C Programming Language,
second edition, Prentice-Hall, 1988.
This is the classic book on C written by its creators.
It is the shortest book I know on the subject and a nice
one to have because of its authorship, but you can also
use other resources to learn C programming for our purposes.
-
Recommended book on Matlab:
Desmond J. Higham and Nicholas J. Higham,
Matlab Guide, third edition, SIAM, 2017.
Webpage of the book
including list of errors.
Matlab's documentation is excellent, but along with its functionality
has reached a scale that requires a lot of sophistication to fully
understand. Moreover, there is a definite role for a book that
is organized by chapter on topics such as all types of functions
(inline, anonymous, etc.), efficient Matlab programming
(vectorization, pre-allocation, etc.), Tips and Tricks, and more.
- Grading rule:
Homework and Quizzes
| Participation
| Class Project
|
45%
| 10%
| 45%
|
-
The homework assignments will be posted.
The homework includes
the computer assignments that are vital to understanding
the course material.
Each homework consists of a written report
that explains what you did and
that responds to the instructions and questions in the problems.
The report should be submitted by e-mail
as well as needs to include your showing your code, its compilation,
and oral discussion of results.
The quizzes will generally be unannounced and brief and
will include the use of learning groups formed by the instructor.
For instance, they may be designed to initiate class discussion
or to give me feedback on your learning.
They may be technical or non-technical in nature.
-
The graded participation component rewards
your professional behavior and active involvement
in all aspects of the course.
Examples of expected professional behavior include
attending class regularly,
reading assigned material when requested,
cooperating with formal issues such as
submitting requested material on time, and
participating constructively in class, specifically in group work.
In this course, professional behavior also includes
adhering to good user behaviors on the shared
computing facilities that you will work on.
-
It is increasingly important
at this point in your education to learn
how to work on a larger project on your own
(with guidance by the instructor)
and to present your results in the form of a
professional-grade type-set report.
The class project will include these components.
You can choose between two types of projects:
One is a topic that is designed individually
and also includes an oral presentation;
this is an ideal way to explore a topic of your choosing
or to start a research project.
The other option is to complete the report
on an assigned class project (without oral presentation).
Additional details or changes will be announced as necessary.
Announcements may be made in class or by e-mail.
You are responsible for checking
your e-mail address sufficiently frequently.
Course Description
Parallel computing has become an ubiquitous
way to perform computer simulations involving large amounts of data or
intensive calculations. The basic purpose of using several processors
is to speed up computations of large problems by distributing the
work. But large problems typically involve vast quantities of data
as well; by pooling the memory from several processors, problems
of previously unsolvable size can now be tackled in reasonable time.
This course will introduce the basic aspects of parallel programming
and the algorithmic considerations involved in designing scalable
parallel numerical methods.
The programming will use MPI (Message Passing Interface),
the most common library of parallel communication commands
for distributed-memory clusters with multi-core CPUs.
Several application examples will
demonstrate how parallel computing can be
used to solve large problems in practice.
We will also consider the options for taking advantage of
shared memory architectures by OpenMP multi-threading and consider using
state-of-the-art massively parallel GPUs (graphics processing units)
as accelerators.
Registered students in this course will gain access to the
cluster maya in the UMBC High Performance Computing Facility
(HPCF; hpcf.umbc.edu).
This class is also intended to familiarize you with this cluster,
if you expect to use it for your research in the future.
One of the side benefits of this class is to help in the creation
of a user community on campus.
The class will include an efficient introduction to the Linux operating
system as installed on the cluster used, and it will include
a review of serial programming in the source code language C
that is integrated into the initial presentation of sample codes.
This review assumes some experience with compiling and debugging
in a high-level source code programming language.
It will only include a restricted set of features of C, but these are
selected to be sufficient for work on the homework assignments in the class.
Learning Goals
By the end of this course, you should:
-
understand and remember the key ideas, concepts, definitions,
and theorems of the subject.
Examples include understanding the purpose of parallel computing
and why it can work, being aware of potential limitations,
and knowing the major types of hardware available.
This information will be communicated in class and in the
textbook, but also in additional reading.
-- This information will be discussed in the lecture as well as
in the textbook and other assigned reading.
-
have experience writing code for a Linux cluster using MPI in C, C++,
and/or Fortran that correctly solves problems in scientific computing.
The sample problems are taken from mathematics and your code has to
compile without error or warning, run without error,
and give mathematically correct results first of all.
In addition, it needs to run on a Linux cluster without error
and you need to be able to explain its scalability, i.e.,
why or why not it executes faster on several processors than in serial.
We will have problems stated in different ways and from various
sources to provide you with exposure to as many issues as possible.
-- This is the main purpose of the homework and most
learning will take place here.
-
have gained proficiency in delivering code written by you to others
for compilation and use.
This includes the concept of providing a README file that gives
instructions how to compile and run the code as well as
of providing a sample output file to allow the user to check the results.
We will work together in class to discuss best practices to transfer code
for homework problems of increasing complexity.
-- You will submit your homework code by e-mail to the instructor
and it needs to compile and run in parallel for credit; this is
complemented by a report that shows and explains your results.
-
have some experience how to learn information from a research paper
and to discuss it with peers.
Group work requiring communication for effective collaboration
with peers and supervisors is a vital professional skill,
and the development of professional skills is a declared learning goal
of this course.
-- I will supply some research papers carefully
selected for their readability and relevance to the course.
Learning from research papers is a crucial skill to develop.
-
have experience with independent work and presenting it both
in a written report and in an oral presentation.
It is vital to gain experience with setting your own goals,
estimating a realistic time line, working with peers and supervisors
on regular updates and giving and receiving suggestions,
submitting and editing a written report to standards of a research paper,
and presenting your results in an oral presentation as part of a
conference session.
-- The class project will cover all of these components,
and we will discuss the various steps necessary in class.
Reading the research papers above will give you a frame of reference
for how the report should look like.
Other Information
Copyright © 2001-2019 by Matthias K. Gobbert. All Rights Reserved.
This page version 1.0, June 2019.