[petsc-dev] controlling vector values while doing matrix free operations

[Dmitry Karpeev]

On Wed, Dec 14, 2011 at 9:35 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> On Wed, Dec 14, 2011 at 19:31, Barry Smith <bsmith at mcs.anl.gov> wrote:
>
>> If you are truly near 0 in temperature or pressure and differencing can
>> push you over the limit then I submit it is perfectly reasonable to solve
>> them as VI.
>>
>
> Note that this tends to grow the algebraic systems quite a bit. Usually
> such problems are solved in conservative variables (density, momentum,
> energy), but the more troublesome constraints (pressure, temperature) are
> nonlinear functions of the conservative variables. As currently
> implemented, you would have to solve (density, momentum, energy, pressure,
> temperature) in order to impose those constraints with SNESVI. I wonder if
> there is a way to eliminate these constraints earlier so that the systems
> don't grow so much.
>
The trouble is that the constraints can get "eliminated" only when they
become active.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111214/7848e0a6/attachment.html>