Re: General--LinearSolve and Precision: Impossible to solve 8x8 matrix
- To: mathgroup at smc.vnet.net
- Subject: [mg70631] Re: General--LinearSolve and Precision: Impossible to solve 8x8 matrix
- From: Bill Rowe <readnewsciv at sbcglobal.net>
- Date: Sat, 21 Oct 2006 05:14:37 -0400 (EDT)
On 10/19/06 at 3:21 AM, abdou.oumaima at hotmail.com wrote:
>Greetings,
>wp = 25; (* WorkingPrecision Value *)
>pr = 20; (* SetPrecision Value *)
>( **** Complexe functions ****)
>Dimensions[L]={8,8}.
>Dimensions[Y]={8}.
>L and Y are function of x.
>K = Normal[Series[LinearSolve[L, Y], {x, 0, 1}]]];
>I can not have result for this linear system within a reasonnable
>time and with a good precision. Any suggestions will be very
>welcome. Cheers.
You haven't provided enough detail for anyone to give any more
than a guess as to how you might achieve better results.
What do you mean by "reasonable time and with a good precision"?
What is the nature of the elements for L and Y? Expressions
containing x as a variable? Linear? Exact coefficients?
Is there any special structure to L? Perhaps band limited or
upper triangular?
If I show I can get what I think is good precision in what I see
as reasonable time for some specific 8 x 8 matrix and 8 element
vector, is that useful to you?
--
To reply via email subtract one hundred and four