[firedrake] Few questions

[Andrew McRae]

1)  Does this region line up with cells of your mesh?  If so, my normal way
is to interpolate that expression into a P0 function:

P0 = Function(mesh, "DG", 0)
f = Function(P0)
f.interpolate(Expression("(x[0] >= 0.25 && x[0] <= 0.30 && x[1] >= 0.25 &&
x[1] <= 0.30) ? 1.0 : 0.0"))

and then use f in calculations.

2)  Pass, someone else can answer this

3)  "Yes", but can you be more specific?  SUPG is just modifying the test
function, so my code looked something like

...
u = ...
alpha = ...
phi = TestFunction(V)
phiSUPG = phi + alpha*dot(u, grad(phi))
a = phiSUPG*....*dx + ....

On 31 August 2015 at 22:31, Justin Chang <jychang48 at gmail.com> wrote:

> Hi all,
>
> 1) How do I define a volumetric source/sink for a specific region? For
> example, if I am solving a diffusion equation with homogeneous BCs and have
> a volumetric source f(x,y) such that
>
> if (x[0] >= 0.25 && x[0] <= 0.30 && x[1] >= 0.25 && x[1] <= 0.30):
>   f == 1
> else:
>   f == 0
>
> 2) I want to play around with PETSc's Variational Inequality
> <http://www.mcs.anl.gov/petsc/petsc-dev/docs/manualpages/SNES/SNESVINEWTONSSLS.html>
> solver. In my own PETSc code all I had to do was provide a Jacobian matrix,
> a residual vector, pass in the options -snes_type vinewtonssls (or it might
> be vinewotnrsls), and pass in SNESVINEWTONSSLS
> <http://www.mcs.anl.gov/petsc/petsc-dev/docs/manualpages/SNES/SNESVINEWTONSSLS.html> .
> From my last discussion with some PETSc folks it seems this VI feature is
> relatively now, so I was wondering if that last item can be invoked through
> the version of petsc4py you guys are using?
>
> 3) Is it possible to employ the SUPG formulation for advection-diffusion
> problems in firedrake?
>
> Thanks,
> Justin
>
-------------- next part --------------
HTML attachment scrubbed and removed