[petsc-users] Using MatCreateComposite with petsc4py
Stefano Zampini
stefano.zampini at gmail.com
Tue Dec 3 00:36:49 CST 2024
Eric
the python equivalent of a MATSHELL is MATPYTHON, where you are in control
of the implementation of the matrix vector product and other operations,
see
https://urldefense.us/v3/__https://petsc.org/release/petsc4py/petsc_python_types.html*petsc-python-matrix-type__;Iw!!G_uCfscf7eWS!cBWJhNKIAgapcwHEU8n-DdtvG1b4efLyBq4hz-hD3u12q055vU-37qhdbu7HsyS9_9yUlAmmIPPCBqfdlCsXdy-GASOOnmY$
This should cover your needs, let me know if you need methods that are not
wrapped
Il giorno mar 3 dic 2024 alle ore 09:15 Pierre Jolivet <pierre at joliv.et> ha
scritto:
>
>
> On 2 Dec 2024, at 11:15 PM, Eric Chamberland via petsc-users <
> petsc-users at mcs.anl.gov> wrote:
>
>
>
> Hello,
>
> We were exploring the use of the MatCreateComposite function with petsc4py,
> but we noticed that it doesn't seem to be available in the Python interface
> (https://urldefense.us/v3/__https://petsc.org/release/petsc4py/reference/petsc4py.PETSc.Mat.html__;!!G_uCfscf7eWS!cBWJhNKIAgapcwHEU8n-DdtvG1b4efLyBq4hz-hD3u12q055vU-37qhdbu7HsyS9_9yUlAmmIPPCBqfdlCsXdy-GYsA6br0$
> <https://urldefense.us/v3/__https://petsc.org/release/petsc4py/reference/petsc4py.PETSc.Mat.html__;!!G_uCfscf7eWS!fphemsDPNmxHUr-ybYLt_xWUrP575eKuI4xixwypEzd2g7uc6qMilt8etg13o24gU0vloOVrsHDXwqlgoY1T7jrcQUiv83PrXvNSYlNU$>
> ).
>
> We were wondering if this is due to a technical limitation or if there is
> a better alternative to achieve the same result.
>
> Our goal is to create a matrix representing PTAP without explicitly
> computing the product.
>
> We also considered using MatShell, but it seems that this approach is not
> supported in petsc4py either.
>
> What do you mean?
> You can use either a MatShell or a MatPython in petsc4py out of the box.
>
> Thanks,
> Pierre
>
> On the other hand, we noticed that MatCreateTranspose is implemented, and
> it aligns somewhat with the kind of operation we're aiming for.
>
> Any insights or advice you could provide would be greatly appreciated!
>
> Thank you,
>
> Eric
>
> --
> Eric Chamberland, ing., M. Ing
> Professionnel de recherche
> GIREF/Université Laval
> (418) 656-2131 poste 41 22 42
>
>
--
Stefano
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20241203/27e357f1/attachment.html>
More information about the petsc-users
mailing list