Re: Help: parallel LU decomposition

Article View: sci.physics.computational.fluid-dynamics

Article #15

Back to sci.physics.computational.fluid-dynamics Previous Next

#15

From: Patrick STOCLET - 44272427
Date: Sat, 27 Jan 1996 00:00

16 lines
1214 bytes

hello,

I had a similar problem also with a CM5 
elliptic pde whith coefficients jumps by multigrid .

1) for building the LU decomposition 
in SIMD  i solve it as a non linear problem ( fixed point method will do
if it is done only once )

2) to apply the preconditionner i  use the truncated neumann method 
   to solve for L and U ( in my case 2 iterations are enough ).

if you need more info send me an e-mail 

best regards 
patrick stoclet

Technical Details

Message-ID: <4ednkr$edu@vishnu.jussieu.fr>

Path: arc1-ca.pugleaf.net!unknown!not-for-mail

References: <4ebgtm$l8n@frodo.rutgers.edu>

Show Full Headers

Complete Article Headers

X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 102daf,e4b2570511d4bb92
X-Google-Attributes: gid102daf,public
X-Google-Thread: 106aa4,e4b2570511d4bb92
X-Google-Attributes: gid106aa4,public
X-Google-Thread: 10acec,e4b2570511d4bb92
X-Google-Attributes: gid10acec,public
From: stoclet@aquarium-2.NoSubdomain.NoDomain.CIS.YALE.EDU (Patrick STOCLET - 44272427)
Subject: Re: Help: parallel LU decomposition
Date: 1996/01/27
Message-ID: <4ednkr$edu@vishnu.jussieu.fr>
X-Deja-AN: 136697388
distribution: world
references: <4ebgtm$l8n@frodo.rutgers.edu>
organization: Universites Paris VI/Paris VII - France
keywords: LU, parallel, preconditioning
newsgroups: sci.physics.computational.fluid-dynamics,comp.parallel,fj.comp.parallel

Reply to Parent Message