Newsgroups: comp.parallel
From: Stephen Barnard <steve@megafauna.com>
Subject: Re: distributed memory iterative schemes
Organization: NAS/NASA Ames Research Center
Date: 18 Jun 1997 20:48:48 GMT
Message-ID: <5o9hjg$2pk@server1.ctc.com>

Kelli L. Robinson wrote:
> populous and see what happens.  I am currently parallelizing a serial
> (f77) code which uses a very basic SOR iterative scheme.  Basically
> its a volumetric stencil (dependicies in all three directions).  The
> boundary conditions are explicit so the stencil is straight forward.

It's not so much an issue of distributed memory as it is an issue of
parallelism in general.  SOR, which is a generalization of Gauss-Seidel,
is not "parallel friendly."  Maybe you should try Jacobi iteration
instead.

	Steve Barnard

--
Articles to parallel@ctc.com (Administrative: bigrigg@ctc.com)
Archive: http://www.hensa.ac.uk/parallel/internet/usenet/comp.parallel