> On Dec 21, 2016, at 2:30 PM, Fangbo Wang <[email protected]> wrote:
>
> Hi,
>
> Background:
> I have a global matrix (sparse) which is very large (2 million by million),
> it is generated from local block matrix with size around 30,000 by 30,000
> (also sparse) using stochastic galerkin projection method.
>
> In the global matrix, I have 700 local block matrices which most of them are
> similar. At the end of the day, I only need to save 45 different local block
> matrices which can save a lot of memory than saving the global matrix.
>
> my problem:
> I don't want to mannually create 45 petsc matrices with similar names, for
> example, A1, A2, A3, A4, etc. I want to automatically create these matrices
> and be able to call these matrices according to its name.
Have an array of matrices?
Mat A[45];
for (i=0; i<45; i++) {
MatCreate(PETSC_COMM_SELF,&A[i]);
MatSetType(A[i],...
....
Barry
>
> Any one have experiences on this? Thank you very much.
>
> Fangbo Wang
>
> --
> Fangbo Wang, PhD student
> Stochastic Geomechanics Research Group
> Department of Civil, Structural and Environmental Engineering
> University at Buffalo
> Email: [email protected]
