Incorrect Gradient Calculated

Note, the solution also changes with the number of points in the mesh.

You need to send the simplest possible complete solver.

Garth

On 17 Nov 2011, at 21:05, Charles Cook <email address hidden> wrote:

> Question #179150 on DOLFIN changed:
> https://answers.launchpad.net/dolfin/+question/179150
>
> Charles Cook posted a new comment:
> Note, the solution also changes with the number of points in the mesh.
>
> --
> You received this question notification because you are a member of
> DOLFIN Team, which is an answer contact for DOLFIN.

On 17 November 2011 22:05, Charles Cook
<email address hidden> wrote:
> My UFL file
>
> element = FiniteElement("Lagrange", interval, 4)
> q = Coefficient(element)
> u= TrialFunction(element)
> w = TestFunction(element)
>
> a = inner(u,w)*dx
> L = inner(grad(q),w)*dx

Try just

a = u*w*dx
L = q.dx(0)*w*dx

and see if you get the same results.

Then send a minimal runnable code that reproduces the problem.

It's not unlikely that there is a bug related to intervals
somewhere, I believe that is much less used and thus
less tested than 2d and 3d problems in fenics.

Martin

Unfortunately I get the same result with u.dx(0).

Where should I send a minimum runnable example?

Charles

-----Original Message-----
From: <email address hidden> [mailto:<email address hidden>] On Behalf Of Martin Sandve Alnæs
Sent: Friday, November 18, 2011 10:41 AM
To: <email address hidden>
Subject: Re: [Question #179150]: Incorrect Gradient Calculated

Your question #179150 on DOLFIN changed:
https://answers.launchpad.net/dolfin/+question/179150

Martin Sandve Alnæs proposed the following answer:
On 17 November 2011 22:05, Charles Cook
<email address hidden> wrote:
> My UFL file
>
> element = FiniteElement("Lagrange", interval, 4) q =
> Coefficient(element) u= TrialFunction(element) w =
> TestFunction(element)
>
> a = inner(u,w)*dx
> L = inner(grad(q),w)*dx

Try just

a = u*w*dx
L = q.dx(0)*w*dx

and see if you get the same results.

Then send a minimal runnable code that reproduces the problem.

It's not unlikely that there is a bug related to intervals somewhere, I believe that is much less used and thus less tested than 2d and 3d problems in fenics.

Martin

--
If this answers your question, please go to the following page to let us know that it is solved:
https://answers.launchpad.net/dolfin/+question/179150/+confirm?answer_id=2

If you still need help, you can reply to this email or go to the following page to enter your feedback:
https://answers.launchpad.net/dolfin/+question/179150

You received this question notification because you asked the question.

Also, anyway to get my email out of the message?

I did a fresh install/update of fenics and the problem still occurs with Beta 2.

I believe I have found the cause. The input might be too 'stiff' or 'sharp'.

I observe the same behaviour by:

- Creating an interval mesh with 100 elements (number of elements doesnt matter much)
- Solving from x=0, to x=1
- Setting the input vector = 1 everywhere except at x=0, setting it equal to 0.955.

Solving with the above system then shows the same behavior.

In short, this is a numerical issue and not a bug :)

This was definitely a resolution problem. I distributed the nodes with a tanh distribution and solved the problem.

Thank you!