[petsc-users] Wrong answers from KSPSolve

[Matthew Knepley]

On Tue, Nov 6, 2012 at 12:38 PM, NAN ZHAO <zhaonanavril at gmail.com> wrote:

> Dear all,
>
> I tried to solve a 18 by 18 linear system with KSPsolve, the right hand
> vector and matrix can be found in the file in the attachment. The right
> hand vector has several zeros and the corresponding row in the matrix only
> have values in diagonal, I expected zero at this element but got -1 for the
> answer. I am using the tutorial code
> in  petsc-3.2-p6/src/ksp/ksp/examples/tutorials/ex1.c as an example to
> assemble the matrix and vector and to solve the system. I also include the
> code in the attachment. I am working on a solution now but have run out of
> ideas.
>
> If you have any hint on the potential solution to this problems let me
> know.
>

There are just so many errors here its hard to begin:

  1) Your exact solution u does not match the rhs

  2) I changed b to match u

  3) The solver you chose was not solving the problem. The default works
fine

Here is the fixed up code. Our job is not to debug your code.

   Matt


> Thanks,
> Nan
>



-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20121106/aef73fa8/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ex1.c
Type: text/x-csrc
Size: 12961 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20121106/aef73fa8/attachment.c>