[aspect-devel] Visualizing actual solution

Jonathan Perry-Houts jperryh2 at uoregon.edu
Thu Oct 29 18:39:07 PDT 2015


build_patches just divides every cell, correct? Is there a way to make
the "surface" layer just a regularly spaced grid of points across
\Omega? You only notice the interpolation issue on the big cells so
those are the ones you want to subdivide, if any. Subdividing the small
cells just produces bigger output files, with basically no noticeable
difference in the figures.

I just made a super inefficient (lots of MPI_Reduce's) postprocessor to
do that, and produced this image of Poiseuille flow calculated with a
1-cell mesh.

It's some awkward code, and probably wouldn't generalize very well to
non-cartesian meshes, but I like the idea. I might be the only one who
cares about this though, and I think even my enthusiasm for it is close
to spent so no need to pursue this further if interest is low.

Thanks for the ideas!

-Jonathan

On 10/29/2015 06:13 PM, Wolfgang Bangerth wrote:
> On 10/29/2015 05:23 PM, Timo Heister wrote:
>> One thing I have never tried but thought might be interesting:
>> - output with build_patches(20) into file1.vtu, load into
>> visit/paraview, display as "surface"
>> - output with build_patches(1) into file2.vtu, also load it in,
>> display as "wireframe"
> 
> Right. Alternatively, for the second input file you could just use what
> GridOut produces.
> 
> Best
>  W.
> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: one_cell.png
Type: image/png
Size: 10935 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20151029/96429d5b/attachment-0001.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 538 bytes
Desc: OpenPGP digital signature
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20151029/96429d5b/attachment-0001.sig>


More information about the Aspect-devel mailing list