Re: [deal.II] Is there an advantage/disadvantage/difference to using FEInterfaceValues (tutorial step 12) instead of the MeshWorker framework (tutorial step 12b) when it comes to DG?

2020-06-30 Thread Wolfgang Bangerth



I would like to implement a DG solver that would solve the 
convection-diffusion equation and later implement another code to solve the 
incompressible NS equation using DG as well. Tutorials 12 and 12b use 
different approaches to solve the pure convection equation, and so I was 
wondering which of these two approaches is better suited for my problem?
The two approaches were developed for different purposes: FEInterfaceValues as 
an approach to compute jump terms for DG methods in the same way as use 
FEValues and FEFaceValues for evaluating the shape functions on cells and 
faces. On the other hand, MeshWorker was meant as a framework that would make 
the assembly of linear and bilinear forms over cells and interfaces easier. If 
FEInterfaceValues had been around at the time MeshWorker was written, it would 
have been one of the building blocks used there.


So there is no real equivalency between the two. The equivalence is more 
between MeshWorker framework and the MeshWorker::mesh_loop() method used in 
step-12 and step-47: As has been expressed many times on this mailing list, 
MeshWorker has a good underlying idea, but it is not well documented or tested 
and none of the people who frequently answer on the mailing list know it well. 
In other words, its higher level functionality is in essence unmaintained. On 
the other hand, MeshWorker::mesh_loop() *is* well documented and supported, 
and has an interface that is relatively widely understood here. My preference 
would have been to just remove step-12b when the previous version was 
rewritten into what step-12 is now, rather than renaming it step-12b. (And do 
the same for step-16b.) I consider step-12b and step-16b as obsolete.


Best
 W.

--

Wolfgang Bangerth  email: bange...@colostate.edu
   www: http://www.math.colostate.edu/~bangerth/

--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/dealii/cf5ffbb2-485e-7deb-f5f7-8633af2c1c35%40colostate.edu.


[deal.II] Is there an advantage/disadvantage/difference to using FEInterfaceValues (tutorial step 12) instead of the MeshWorker framework (tutorial step 12b) when it comes to DG?

2020-06-30 Thread CopyCat
Hello,

I am in the process of learning DG and dealii. 
I would like to implement a DG solver that would solve the 
convection-diffusion equation and later implement another code to solve the 
incompressible NS equation using DG as well. Tutorials 12 and 12b use 
different approaches to solve the pure convection equation, and so I was 
wondering which of these two approaches is better suited for my problem? 

Thanks 

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/dealii/c127437d-778e-4a1a-a5b0-72b5c4c4b51co%40googlegroups.com.