Re: Large sparse system
Have you tried reordering the matrix (using RCM/COLMD, etc) to reduce
the fill-ins during factorization?
The technique you ultimately use depends on the nature of the matrix
(which at this point is unknown to me). If is SPD, then you could
perhaps compute its cholesky factor and do only a partial back-solve. I
wouldn't necessarily recommend an iterative solver unless you can
somehow verify that the matrix is reasonably well-conditioned (or can
be cheaply preconditioned) and/or a matrix-vector product can be formed
without having to explicitly compute the entries of A.
.
Relevant Pages
- Re: [SI] Convergence - Frank ess Comments
... Jim Kramer - http://www.pbase.com/shootin/image/58113527 ... I like wood, I like nature colors, and I cannot not smile when I see one; relieved to see Thomas is accepting this mayhem in good cheer; I empathize with the lighting conundrum, poor practitioner that I am, and might have gone a little more "high key" (see! ... This one really strikes a chord with me; as in all art-making machinery, media, and apprehension, there is a plethora of convergences, and this work calls up some of the most poignant; an alternate technique could draw good focus from front to back, perhaps, but this configuration also invokes other characteristics of art appreciation, more fulfilling than a technical _coup_. ...
(rec.photo.equipment.35mm) - Re: Coil failure again?
... Using the patented Mavis Beacon "Hunt&Peck" Technique, ... nature, my bet would be on the black box itself. ... spray and a change of black box - in fact I suspect he didn't bother to switch the black boxes back round after changing them - cos there's a white label on this that I don't recall seeing before. ...
(uk.rec.motorcycles) - Selection of a Suitable AI Technique
... If the nature of computations required in a task is not well understood ... or there are too many exceptions to the rules or known algorithms are ... On the Selection of a Suitable AI Technique for Solving a Given ...
(comp.ai) - a question to Bob Silverman about factorization
... the fusion problem as they have been ignoring the teaching of nature. ... not saying that we cannot do better than nature, ... Are mathematicians making the same mistake about the factorization problem? ...
(sci.crypt) |
|
