[petsc-users] MatMPIBAIJSetPreallocationCSR i and j indices

[Matteo Parsani]

Hi Barry and Jed,
just a curiosity and maybe a good point to learn something:

1- why are those* inconvenient *"convenience routines" ?

2 - why those routines *should not be used* when writing a fresh code or a
large re-factoring?


Thanks in advance.



On Sun, Sep 29, 2013 at 3:52 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> Barry Smith <bsmith at mcs.anl.gov> writes:
>
> >    We can do all this, but how much effort do we really want to make
> >    for supporting these inconvenient "convenience routines"?
>
> Not much, but this is a tiny amount of code.
>
> >    BTW: We should really label these function with
> >
> >    Level: Recommended only for legacy applications being quickly ported
> to PETSc
> >
> >     I really only want people who already have a working non-PETSc
> >     code that uses CSR directly already to use these routines so they
> >     can quickly try out PETSc, I never want someone writing fresh code
> >     or doing a large refactorization to EVER use these routines.
>
> Agreed.
>
> >> On second thought, why don't we just document that
> >> MatMPIBAIJSetPreallocationCSR respects MAT_ROW_ORIENTED within each
> >> block and let the user choose which they want to use?
> >
> >     Also there is no way to get this flag into
> MatCreateMPIBAIJWithArrays()
>
> The user calls MatSetOption before making that call and changes it
> afterward if they want.
>



-- 
Matteo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20130929/cb8916f8/attachment.html>