EAGLE Physics Validation
The current tests of the EAGLE model are completely scattered and we really don't know if all of the physics is correct. We have currently defined success as matching the old EAGLE code. This no longer seems possible, and regardless this old code was not well tested.
This is going to be quite a long issue but I think it will provide us with a full checklist of things that need to be confirmed (self consistently, so even if it has been tested before it should be tested again). If you have previous plots, please feel free to provide them and we will both re-run those tests.
Hydrodynamics
Physics to be checked
-
Conservation of angular momentum -
Diffusion convergence with resolution -
Diffusion does not transport energy against gradients -
Wendland-C2 is the best choice of kernel (minimal h given softening) -
Hydrodynamics is cosmology-independent
Task list
-
Re-name ANARCHY-DU to SPHENIX -
Re-run Santa Barbara and compare Minimal to SPHENIX -
Re-run Nifty cluster and compare against published results -
Show convergence on hydrodynamics problems -
Check energy as a function of time with the Evrard collapse -
Run tests with different kernels and compare results -
Isolated galaxy with different kernels (minimal h larger than typical expected h) and check pairing instability -
Check cosmological versions of tests
Cooling
Physics to be checked
-
Cooling rates are correctly calculated and are applied to particles correctly -
We correctly account for any lost energy when applying cooling in the energy.txt
-
Cosmological corrections for the cooling rates are correct
Task list
-
Check cooling rates v.s. tables (including with cosmology) -
Check energy conservation in tests with cooling (start with a hot box and let it cool).
Star Formation
Physics to be checked
-
We create the correct number of stars given star formation rate?
Task list
-
Validate that star formation rate matches stars formed in a cosmological run
Feedback
Physics to be checked
-
Stars inject the right amount of energy when doing SNII feedback -
Stars enrich the gas correctly, and produce a consistent amount of metals
Task list
-
Check energy injection of stars in an isolated environment (more than one star at once) -
Check enrichment of stars in an isolated environment (more than one star at once)
Black Holes
Physics to be checked
-
Does accretion work? -
Do black holes grow? -
Do black holes merge? -
Do black holes do feedback? -
Do black holes inject the correct amount of energy?
Task list
-
Create isolated black hole tests to address the above questions -
Create isolated galaxies with hot haloes that can check black holes in a cleaner environment
Generic Tasks
During this, we should aim to do the following other things:
-
Fix any used python scripts to no longer require usetex=True