Using min_max_speed_davis instead of min_max_speed_naive as default

[DanielDoehring]

As discussed during the last meeting, I looked into the possibility to replace min_max_speed_naive with min_max_speed_davis.

Except for https://github.com/trixi-framework/Trixi.jl/blob/276dc3ca22a4c72b81aa6659e57ff1382dd1ae0c/examples/paper_self_gravitating_gas_dynamics/elixir_eulergravity_jeans_instability.jl

every example works (except for the polytropic euler equations, for which min_max_speed_davis can be quickly added ) out of the box, i.e., without the need to adjust e.g. CFL numbers.

For Euler-Gravity-Jeans-Instability a reduction in CFL does not resolve the crash of the simulation. At least for this example one would need the naive wave speed estimate. Question is whether one would place it in e.g. the elixir.

[sloede]

For Euler gravity Jeans instability reduction in CFL does not resolve the crash of the simulation. At least for this example one would need the naive wave speed estimate. Question is whether one would place it in e.g. the elixir.

That would be fine for me if @ranocha and @andrewwinters5000 are also ok with it

[ranocha]

👍

[sloede]

@DanielDoehring do you still plan to include this in v0.6? I'm asking since in #1705 we have finished nearly all must have merges, thus I'm pinging the maybes 🙂

[DanielDoehring]

No, this is a bit too large to be finished soon. I aim for doing this (latest) with 0.7 when the two-layer shallow water is moved somewhere else.

[sloede]

OK, I'll take it off the list for #1705 👍

[DanielDoehring]

I set up a draft PR where we would keep the min_max_speed_naive as it might be helpful in some cases (as seen for the jeans instability referred to above): #1743

Question is whether I should in some elixirs (preferably those for which not an abundance of derived tests exist) substitute in the now standard flux_hll.

[ranocha]

I think it's fine to use the new standard flux_hll 👍