[petsc-users] da_overlap vs width in dmdacreate

Matthew Knepley knepley at gmail.com
Sat Jun 25 20:06:11 CDT 2016


On Sat, Jun 25, 2016 at 8:03 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:

>
>  On Jun 25, 2016, at 7:56 PM, Matthew Knepley <knepley at gmail.com> wrote:
> >
> >
> > On Sat, Jun 25, 2016 at 7:18 PM, Xiangdong <epscodes at gmail.com> wrote:
> > Maybe my previous email is not clear. What is the relationship between
> the parameter x in DMDASetOverlap and the parameter s in DMDACreate2d? Does
> x need to be smaller than s? Are they completely independent?
> >
> > Yes. The Overlaps are only used in decomposing the DA into small blocks
> for things like block GS and NASM.
>
> The manual page for DMDASetOverlap() is pitiful, it could at least
> indicate what this overlap thing is. Even the sentence above has far more
> information than the manual page but is still completely unclear.
>

I agree with you. However, I think most of what Peter did here should be
canned. I am trying to rewrite
SNES completely to take these things into account (he ignores BCs most of
the time, does not interface
well with unstructured or SAMR, etc.), but it got bogged down just like me.

   Matt


> >
> >  Thanks,
> >
> >     Matt
> >
> > PetscErrorCode  DMDASetOverlap(DM da,PetscInt x,PetscInt y,PetscInt z)
> >
> > PetscErrorCode  DMDACreate2d(MPI_Comm comm,DMBoundaryType
> bx,DMBoundaryType by,DMDAStencilType stencil_type, PetscInt M,PetscInt
> N,PetscInt m,PetscInt n,PetscInt dof,PetscInt s,const PetscInt lx[],const
> PetscInt ly[],DM *da)
> >
> > Thanks.
> >
> > Xiangdong
> >
> >
> > On Fri, Jun 24, 2016 at 5:23 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> >
> >    What arguments are you referring to?
> >
> > PetscErrorCode  DMDACreate2d(MPI_Comm comm,DMBoundaryType
> bx,DMBoundaryType by,DMDAStencilType stencil_type,
> >                           PetscInt M,PetscInt N,PetscInt m,PetscInt
> n,PetscInt dof,PetscInt s,const PetscInt lx[],const PetscInt ly[],DM *da)
> >
> > the argument s is the stencil width and depends on the finite difference
> stencil you want to use. There is no da_overlap argument.
> >
> >   Barry
> >
> > > On Jun 24, 2016, at 4:17 PM, Xiangdong <epscodes at gmail.com> wrote:
> > >
> > > Hello everyone,
> > >
> > > I have a quick question. What is the relationship between da_overlap
> and the width parameter in dmdacreate2d?
> > >
> > > From the examples, it seems that width=1 works with da_overlap=4.
> Could you please explain the relationship or point me to a reference page?
> > >
> > > Thanks.
> > >
> > > Xiangdong
> >
> >
> >
> >
> >
> > --
> > What most experimenters take for granted before they begin their
> experiments is infinitely more interesting than any results to which their
> experiments lead.
> > -- Norbert Wiener
>
>


-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20160625/d69fb700/attachment-0001.html>


More information about the petsc-users mailing list