Newsgroups: comp.parallel,comp.parallel.mpi
From: ghill@bigtime.cs.colorado.edu (Gregory Hill)
Subject: MPI-RGL 0.4
Summary: third test flare by temporary mod.
Organization: University of Colorado, Boulder
Date: 2 Aug 1996 22:14:06 GMT
Message-ID: <4ttuje$kjp@onramp.arc.nasa.gov>

Followup-To: comp.parallel.mpi

We have implemented an extension to MPI called MPI-RGL: a Regular Grid 
Library for MPI. This interface provides a usability layer for the application 
specialist who wishes to parallelize a code whose data structures are
regular grids. For this the details of the underlying message
passing library, in this case MPI, are immaterial -- the important
concepts are domain decomposition, exchange of information across
boundaries and loop index modification. We provide simple functions
which take care of this. In addition we propose a simple interface 
for parallel I/O. We have also provided simple, easy-to-use profiling 
and debugging capabilities.

We have just released a major new version of MPI-RGL: 0.4.

To get it please use the form on the home page:
   http://www.cs.colorado.edu/homes/clive/public_html/MPI-RGL.html

We have made numerous changes: tiding up the code, fixing bugs, etc,
plus we have added easy-to-use profiling and debugging capabilities.

We have also made the code self-documenting, using this feature to
generate man pages in nroff, latex and HTML format.
The manual is fully updated and on the Web as Postscript and HTML.

Please update to the new version; send any problems to
   rgl-bugs@cs.colorado.edu.

The RGL development team.
-- 
|          O___|     |                              |
| Montani  /\  \_    |   Sentiment without action   |
| Semper   \ \/\_\   |   is ruination of the soul.  |
| Liberi    U \_  |  |    - Edward Abbey            |
			[ Cute Greg, cute  --mod.]

