[petsc-dev] forget about libmpiuni completely
Barry Smith
bsmith at mcs.anl.gov
Wed Feb 17 16:54:57 CST 2010
We want both the transparency and simplicity to the user PLUS a
common way of handling things in PETSc to reduce the learning curve of
PETSc developers and PETSc maintenance.
Each "special case" in PETSc is a flaw in our design, and an
admission we made a mistake or didn't really know how to do things.
Currently the MPIuni stuff is a bit of special case.
Barry
On Feb 17, 2010, at 4:48 PM, Matthew Knepley wrote:
> On Wed, Feb 17, 2010 at 2:44 PM, Lisandro Dalcin <dalcinl at gmail.com>
> wrote:
> On 17 February 2010 12:00, Matthew Knepley <knepley at gmail.com> wrote:
> > I am for a separate mpiuni that get downloaded. I think this makes
> more
> > sense with
> > our other packages concepts, and is more modular. We can reuse so
> much more
> > of
> > configure as well.
>
> I still fail to see the pros of having a separate mpiuni.
>
> I think you have to put on user's side when taking a decision about
> this. Are FINAL USERS going to have any benefits for having mpiuni as
> a separate download and having to pass a separate library in when
> invoking the linker? If a user wants no-MPI, Is there any other
>
> They do not pass anything. The beauty of the Petsc build.
>
> fake-mpi alternative supported in PETSc? If the answer is no, so then
> why bother people with a separate download?
>
> They do not even have to download it. They specify the same option
> as before.
> From a user perspective, it does not change. From a Petsc and
> library developer,
> it is better.
>
> Matt
>
> --
> Lisandro Dalcin
> ---------------
> Centro Internacional de Métodos Computacionales en Ingeniería (CIMEC)
> Instituto de Desarrollo Tecnológico para la Industria Química (INTEC)
> Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET)
> PTLC - Güemes 3450, (3000) Santa Fe, Argentina
> Tel/Fax: +54-(0)342-451.1594
>
>
>
> --
> 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
More information about the petsc-dev
mailing list