preconditioner for matrix-free linear system

[Yongcheng Zhou]

Thanks Matt, Satish and Barry. I think I have to convert my matrix into a PETSc
matrix in block AIJ type. I wan reluctant to do this since I might need to solve
linear system many times and I had preferred to use matrix free approach.

Another question. In calling PETSc for solving the linear system for the
second time, I got this:

  You can not initialize PETSc a second time!

But I did destroy and finalized last PETSc structure, via

  MatDestroy(A);
  VecDestroy(x);
  VecDestroy(b);
  KSPDestroy(ksp);
  PetscFinalize();

Any suggestion is greatly welcome!

Rocky




On Tue, Apr 14, 2009 at 12:28 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
>
> "PETSc's powerful preconditioners" are constructed from knowledge about the
> matrix entries; if you do not provide any information about the matrix
> entries it cannot
> construct any preconditioners. You need to provide the matrix as a PETSc
>  AIJ or BAIJ Mat
> in order to use those preconditioners.
>
>  Barry
>
>
>
> On Apr 14, 2009, at 1:04 PM, Yongcheng Zhou wrote:
>
>> hi there,
>>
>> I want to link my own package with PETSc in order to make use of its
>> various preconditioners.  I am using
>> matrix free method, so that I can directly refer the large matrix
>> saved in my own format. The connection
>> works OK without preconditioners, but runs into trouble when most of
>> the preconditioners is called.
>> For example, I got this message when using PCICC preconditionder:
>>
>> Matrix format shell does not have a built-in PETSc direct solver!
>>
>> So my question is how to utilize PETSc's powerful preconditioners for
>> my matrix-free application.
>>
>> Thanks,
>>
>> Rocky
>
>