On 10/27/05, Till Isambert <address@hidden> wrote:
I'm solving a (large) linear system of equations with a LU decomposition
in GSL. I'm wondering if it would be faster to solve it with a
preconditionned conjugate gradients method ? If such, which function
should i use ..? Sorry if my question is trivial, actually i've looked
inside the GSL documentation but i'm not so sure about how to proceed.
You can certainly use a conjugate gradient method. The standard method
seems to be the one due to Hestenes and Stiefel (Magnus R. Hestenes
and Eduard Stiefel, 1952, Methods of Conjugate Gradients for Solving
Linear Systems, Journal of Research of the National Bureau of
Standards 49, 409-436). One of the advantages is that it can be made
to work with sparse matrices. It's not part of the GSL, but perhaps
you can use or adapt the CG minimizers that are included.
-- mj