 |
MGNet |
|
|
|
|
|
Introduction |
|
|
This is a repository for information related to multigrid, multilevel,
multiscale, aggregation, defect correction, and domain decomposition methods.
These methods are used primarily by scientists and engineers to solve
partial differential equations on serial or parallel computers. |
|
|
Multigrid has the property of using linear time and space to solve a
collection of interesting problems, thus making it a very fast, robust
solver.
Domain decomposition methods are extremely useful for solving problems on
oddly shaped domains or for parallelizing standard iterative methods. |
|
|
While MGNet tries to be the first and last place anyone interested in
multigrid or domain decomposition methods needs to look, this is obviously not
always possible.
Here
are some other sources of interest to the multigrid or domain decomposition
communities. |
|
|
You may retrieve any file in MGNet.
You may also submit files to the archive. |
|
|
|
|
|
Warning:
Unless stated otherwise, everything in this repository is use at your own
risk.
There are no warantees implied or expressed, not even that anything does what
it claims to do or will even install. |
|
|
|
|
|
Newsletters |
|
|
At the beginning of each month, a newsletter is sent out to the entire MGNet
mailing list.
To post a message to the MGNet newsletter, send electronic mail to
The newsletter always needs more content.
Items that are particularly sought after include
technical discussions (but not tirades),
meeting and course information relevant to multigrid,
and
abstracts for codes, papers, and books. |
|
|
You can receive a free subscription to the newsletter
by sending a message to
with your name and electronic mail address in the message.
You will also receive more information about MGNet when your request is
processed. |
|
|
|
|
|
Preprints |
|
|
MGNet contains a large number of papers on many topics, not just multigrid.
However, most papers have a partial differential equation or two in them and
are aimed at an audience of scientists and engineers. |
|
|
|
|
|
Bibliography |
|
|
MGNet's bibliography grows whenever new entries come to the attention of its
authors. (Please send yours.)
This is a large BibTeX database which you are free to use.
The bibliography is available as a document, too. |
|
|
|
|
|
Tutorials |
|
|
If you are new to the multigrid or domain decomposition fields, or just wonder
what MGNet is all about, look here.
There are pointers to some rather nice tutorials on the subject. |
|
|
|
|
|
Free Software |
|
|
A number of software packages are stored on MGNet.
Some are public domain, some are copyrighted, and some may someday be
copyrighted.
If you have a multigrid, domain decomposition, or parallel code or package
you can place a copy in MGNet, too. |
|
|
|
|
|
Virtual Proceedings |
|
|
Several multigrid conferences have a habit of publishing pre-proceedings or
electronic proceedings.
The Copper Mountain (Colorado, USA) multigrid conferences (held on odd
numbered years in April when the snow is still good for skiing) and the
GAMM workshops on parallel multigrid (held in Germany and Austria from time
to time at rather nice places) are examples of this. |
|
|
Find out why multigrid and domain decomposition are such nice fields to
work in. |
|
|
|
|
|
Conference Announcements |
|
|
Numerous conferences are planned, months in advance.
Here are a some that would interest MGNet's audience.
This area of MGNet changes often. |
|
|
|
|
|
Mystery Link |
|
|
This changes from time to time. usually on Monday.
The goal is for it to change once a week, but monthly seems to be the
frequency at the moment. |
|
|
|
|
|
The Sponsors |
|
|
MGNet is supported in part by
Yale University,
Department of Computer Science,
New Haven, Connecticut, USA.
CERFACS,
Parallel Algorithms Group,
Toulouse, FRANCE.
University of Kentucky,
Department of Mathematics and the
Center for Computational Sciences,
Lexington, Kentucky, USA.
IBM's Research Division was a sponsor from 1991-1996.
In addition, many people have contributed suggestions and their own time to
make MGNet a success.
If you have a suggestion, please send me a note. |
|
|
|
|
|
Cheers, |
|
|
Craig C. Douglas |
|