milonga profiling test examples

[Vitor Vasconcelos]

Hi guys,

Riding on the message of César, I would like to hear from you about one thing.

I'm (very very slowly) working on how to parallelize milonga. Yesterday I was
thinking on good examples to run in order to profile milonga and check
what are the most costly functions. In order to do that I came up with
four tests.

Based on the test-3dshape provided by milonga itself (thanks Germán) I
changed few lines
in the geo file and also generated a hexahedral mesh.

With a tet and a hex mesh I'm planning to run milonga for elements and volumes.

So, in total, 4 cases. All of it solving with s2 method.

No big changes on the source code yet, I'm justing taking notes of my
ideas (putting everything on the paper) in order to think on the
problem before work on it.

My big question: is this representative enough to have a clue on what
are the most important candidates for parallelization? Do you have
suggestions? Should I profile diffusion too?

That's it. Thanks!

Vitor

[Jeremy Theler]

This sounds great Vitor!

I would add diffusion too. It is a completely different formulation.

Also you might want to try 1st and 2nd order elements.

Do you know the story of the chess checkerboard, right? After a n ideas one ends up with 2^n runs.

Use the 3dshape but stick to geometries with analytical solutions, i.e. sphere, cylinder or prims.

Then you can use the bunny.

My candidates for paralellization are the routines that compute the elemental matrices (both in diffusion and SN) in FEM and the routines that find neighbors and the ones that build the big matrices in FVM.

[Vitor Vasconcelos]

Great to hear from the father of the child. ;-)

I'll follow your ideas and, yes, your candidates are pretty much the
costly functions accordingly to my first naïve tests.
Anyway, I need to do it with method before going further.

I'll keep you all informed of any progress.

Regards,

Vitor

> --
> You received this message because you are subscribed to the Google Groups
> "wasora" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to wasora+un...@seamplex.com.
> To post to this group, send email to was...@seamplex.com.
> Visit this group at https://groups.google.com/a/seamplex.com/group/wasora/.
> To view this discussion on the web visit
> https://groups.google.com/a/seamplex.com/d/msgid/wasora/ea4f79639a7bb829367698519d682239de910056.camel%40seamplex.com.
> For more options, visit https://groups.google.com/a/seamplex.com/d/optout.