Re: solving linear equations with LAPACK/Matlab
- From: spellucci@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx (Peter Spellucci)
- Date: Mon, 19 Dec 2005 11:59:18 +0000 (UTC)
In article <1134824703.183866.236370@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
wakun@xxxxxxxxx writes:
>Hi there,
> I am solving linear equations with Matlab and now I need to use C++
>instead. The size of the matrix is about 200x200 and is invertible. In
>matlab, I solve the problem as follow
>
>%Ax = b the system
>x=inv(A)*b
no !
x=A\b;
>
>(BTW, is it a right way to solve the system in matlab? Any better
>routine for that problem?)
>
>Anyway, for some reason, my tutor ask me to solve the same problem in
>C++. So I am going to solve it with the help of LAPACK and the C++
>wrapper (LAPACK++). However, I am very new to LAPACK and the
>documentation seems not very friendly to user. I guess I should find
>the routine to calculate the inverse of the matrix first and multiply
>the inverse matrix with the vector on RHS (b). Would you please tell me
>if I am doing right, any function in LAPACK for solving such system
>directly ? Any example for solving tridiagonal system, five-diagonal
>system and seven-diagonal system ?
>
>Thanks in advance.
>
(already done). you wonder why it works? .. because your system
administrator has installed the library which contains it all
hth
peter
.
- References:
- solving linear equations with LAPACK/Matlab
- From: wakun
- solving linear equations with LAPACK/Matlab
- Prev by Date: Re: Least squares curve fitting
- Next by Date: Basic question about Neumann boundary conditions
- Previous by thread: Re: solving linear equations with LAPACK/Matlab
- Next by thread: Order of Magnitute(Big Oh)
- Index(es):
Relevant Pages
|
