Boyce,
I've added the MPIAIJ version of MatAXPY to petsc-dev which does the
preallocation before doing the addition. Please test it with your code and
report any problems if you have.
Thanks,
Shri
----- Shri <abhyshr at mcs.anl.gov> wrote:
>
> ----- Boyce Griffith <griffith at cims.nyu.edu> wrote:
> >
> >
> > On 9/27/10 5:28 PM, Shri wrote:
> > > Barry,
> > > I've modified MatAXPY_SeqAIJ as follows when
> > > DIFFERENT_NONZERO_STRUCTURE flag is true (i) create a new matrix and do
> > > preallocation for it. For determining the number of nonzeros per row in
> > > the new matrix, i've used the number of nonzeros in the original two
> > > matrices. This is the piece of code which computes the nnz per row in the
> > > new matrix
> > > for(i = 0;i< m;i++) {
> > > nzx = xi[i+1] - xi[i]; nzy = yi[i+1] - yi[i];
> > > if (nzx> nzy) nnz[i] = nzy + (nzx - nzy);
> > > else nnz[i] = nzx + (nzy - nzx);
> > > }
> >
> > I don't understand why this is the right thing to do.
> My apologies,the above code snippet is incorrect. Jed has already pushed a
> fix for this and i'm working on it too. Sorry about the confusion.
>
> Shri
> It seems like
> > this either sets nnz in the new matrix to be nzx or nzy, whichever is
> > smaller. Am I misreading this code snippet?
> >
> > -- Boyce
>
