Added example and compilation instructions for Insitu #1661
Conversation
| ## Simulation of a lid-driven cavity | ||
| In this case we simulate a lid-driven cavity with smoothened belt velocity to fulfil the continuity equaiton in the corners. There is a 3D example (with 3 elements in the spanwise $z$ direction) and a pseudo-2D example (with a 2D mesh that then generates a case with one element in the spanwise direction). | ||
|
|
||
| The `lid.box` file needs to be converted to a .re2 file using the Nek5000 tool genbox, and then using rea2nbin into the `lid.nmsh` Neko mesh file. |
There was a problem hiding this comment.
Looks good! Would it be possible to use genmeshbox instead? To avoid using rea2nbin.
Would it also be possible to provide a description of what the insitu part does?
| @@ -0,0 +1,4 @@ | |||
| #!/usr/bin/bash | |||
|
I think the name is fine, as long as we have some description in the readme |
| @@ -6,6 +6,7 @@ libdir=${exec_prefix}/lib | |||
| includedir_pkg=${prefix}/include/neko | |||
| FC=@FC@ | |||
| FCFLAGS='@FCFLAGS@' | |||
| extra_files=@adios_files@ | |||
There was a problem hiding this comment.
Not sure I understand why it doesn't work without this. The object file should already be in libneko from the main Makefile.am
Lines 482 to 510 in 18e7a84
Maybe the ax_adios.m4 failed to detect it? (but then it shouldn't have been built hmm
|
Just a quick note about the name, as I currently don't have time to properly review. I would simply update the existing lid example. It does not demonstrate much new neko concepts compared to other examples in its current form, and this is an opportunity to make it unique. Kind of echoing your comment from before @MartinKarp that it is better to have a few examples, but richer. |
|
Hi, I agree with your comments. I will proceed with them. |
Hi,
I added a new variable to makeneko to include the ADIOS files during compilation. I’m not sure if this is the best approach. In the future we might have streaming as a simulation component, so it would be good if that file is taken into account during the linking of Neko itself (Maybe it already is). I’m happy to adjust this if needed—this is just one possible solution.
I’ve also added an example that demonstrates starting a simulation in Neko, transferring data to Python, and writing from Python. Do you think there’s a better way to name this example? I avoided integrating it directly into the lid example because requiring ADIOS to run it might be overkill.