fenics team mailing list archive

["Theodore Omtzigt" <theo@xxxxxxxxxxxxxxxxx>]

When you guys talk about solvers are you talking about the PETSc/Trilinos
level or are you talking about a solver scheme such as SIMPLE/PISO, etc?

If you are talking about PETSc, I think it would be good to define an
interface that the solver needs to be able to respond to so that lib
dependency management is trivialized. If it is in the form of solver
schemes, it would be good to define a data structure + operator generic
programming interface so that the infrastructure is being build for all
solvers. I have pointed out this before but the generic programming folks
have made some good progress on this. BOOST with MTL, and domain specific
efforts like DUNE, and GrAL. 

Theo

-----Original Message-----
From: fenics-dev-bounces@xxxxxxxxxx [mailto:fenics-dev-bounces@xxxxxxxxxx]
On Behalf Of Johan Hoffman
Sent: Wednesday, January 09, 2008 7:29 AM
To: Discussion of FEniCS development; Claes Johnson
Subject: Re: [FEniCS-dev] Solvers

Hi all,

This sounds good: anyone that wants to put a solver under fenics.org
should be encouraged to do so. It has many benefits. Any conditions for
doing so should be simple; I suggest that all that is required is that the
solver project fits with the FEniCS vision stated at the homepage, just as
for the other projects, and as Anders says that there is a homepage with
depencies etc. There will then be a natural selection among the solvers,
if they are not used they will probably tend to fade away by themselves
(just like the other projects at fenics.org).

As you say, we are now in the process of publishing our solver project
Unicorn for continuum mechanics under fenics.org, which is expected to
happen this week or the next. A webpage is already in place (which we will
brush up in time for the release): http://www.fenics.org/wiki/Unicorn

This project is based on the suite FIAT/FFC/DOLFIN, and more info will be
available shortly.

The reasonable thing is to present this project along with all other
FEniCS projects at the main page. But of course, if the number of solvers
grow too large we may consider dividing projects at fenics.org into
subsets, such as form compilers, PSEs, pre- and postprocessing, etc. But
at this point I do not see any reason for doing so.

/Johan


> Hi all,
>
> We have discussed a few times before how to organize a collection of
> solvers on fenics.org (with limited success). One obstacle we had was
> that no one was willing to maintain the solvers, but this does not
> mean that we can't find (another) good solution.
>
> My suggestion would be that we set up a (very loose) framework for
> solvers based on FEniCS (in particular DOLFIN). We can have some
> common requirements for solvers, like having a web page on the wiki,
> needing to state which version of the other packages it depends on,
> having a compatible license (GPL or LGPL) etc. The point would be to
> make it easy to add solvers. We would have a common page for all
> solvers, which lists the available solvers and link to that page from
> the main page.
>
> If this sounds reasonable, we can discuss exactly which (few)
> requirements we want to make and get this going soon. I know that
> Johan^2 have an FSI solver they can add, and Kristian has a plasticity
> solver that can be added. Others may follow.
>
> --
> Anders
> _______________________________________________
> FEniCS-dev mailing list
> FEniCS-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/fenics-dev
>


_______________________________________________
FEniCS-dev mailing list
FEniCS-dev@xxxxxxxxxx
http://www.fenics.org/mailman/listinfo/fenics-dev