[petsc-dev] Possible Givens Project

[Fande Kong]

On Mon, Jan 25, 2016 at 11:00 AM, <petsc-dev-request at mcs.anl.gov> wrote:

> Send petsc-dev mailing list submissions to
>         petsc-dev at mcs.anl.gov
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://lists.mcs.anl.gov/mailman/listinfo/petsc-dev
> or, via email, send a message with subject or body 'help' to
>         petsc-dev-request at mcs.anl.gov
>
> You can reach the person managing the list at
>         petsc-dev-owner at mcs.anl.gov
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of petsc-dev digest..."
>
>
> Today's Topics:
>
>    1.  Possible Givens Project (Matthew Knepley)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sun, 24 Jan 2016 20:34:07 -0600
> From: Matthew Knepley <knepley at gmail.com>
> To: PETSc <petsc-dev at mcs.anl.gov>
> Subject: [petsc-dev] Possible Givens Project
> Message-ID:
>         <CAMYG4Gk1z_KGg=cgmfxzakRmO65hE3PqbNc=SURnPhnvQk5Y=
> g at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Some parts of SNES need real help.
>
> 1) NGMRES is purely serial
>
> This just requires a parallel TSQR. We could just write it, since I was
> trying to avoid
> a hard dependency on Elemental until Jack has a Damascene conversion and
> rewrites
> it in C using regular make, but that is a game-time decision.
>
> 2) NASM needs to really work
>
> It needs to be rewritten, generalized, and streamlined so that it can do
> everything that
> student of Xiao-Chuan are currently doing by hand. This seems like the most
> important
> thing to me.
>

I think we possibly also need to support NASM at a pure algebraic level.
ASPIN requires us explicitly set the domain decomposition, but in fact this
information could be extracted from the matrix and vector layouts.



>
> 3) The internals are a mess
>
> I have a branch knepley/snes-rewrite that completely reworks the internals
> and I think
> makes them much easier to understand and maintain. This project would
> mostly be
> getting the tests in a state that switching over made sense.
>
>    Matt
>
> --
> 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-dev/attachments/20160124/621975c1/attachment-0001.html
> >
>
> ------------------------------
>
> _______________________________________________
> petsc-dev mailing list
> petsc-dev at mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/petsc-dev
>
>
> End of petsc-dev Digest, Vol 85, Issue 16
> *****************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20160125/77f045c9/attachment.html>