Martin Kronbichler
The proposed changes are applying a tool mentioned in #104 (`VectorTools::point_values`), but they do not address the code in question. The statistics calculators could probably use something similar, but I...
Regarding the flow past cylinder case: I ran it with 1 level of refinement and degree 4, and got 6 changes in the last digit for the whole output file....
> To justify the changes of the code in LinePlotCalculator (reducing the modularity of the code) by performance improvements, I think one needs to argue (i) which performance improvement comes...
Regarding the observed instability: This seems to be cured by using a ramping in the start of the simulation, as going to full speed for high Reynolds numbers is obviously...
In 3449da6 I added better comments and re-organized the code: I now first compute the surface areas and then go through the cells, which allows me to directly store the...
I can make those changes, but before we spend more time on this side topic, let me raise a question: What would you think if we introduce a special parameter...
In the constructor of `FEEvaluation` we can only afford a reasonably cheap operation. A solution actually looking into the quadrature formulas and determining whether it is the most accurate would...
> @DanielPaukner Could you give a try to this code. I am not sure what kind of application you are considering, but I anticipate that typical beams will need to...
@ivweber and I went through the code and re-based. I think most of the comments were addressed, and I saw some things that we might want to fix some follow-up...