You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+3-2Lines changed: 3 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -33,6 +33,7 @@ import hmcode
33
33
# Ranges
34
34
k = np.logspace(-3, 1, 100) # Wavenumbers [h/Mpc]
35
35
zs = [3., 2., 1., 0.5, 0.] # Redshifts
36
+
T_AGN = 10**7.8 # Feedback temperature [K]
36
37
37
38
# Run CAMB
38
39
parameters = camb.CAMBparams(WantCls=False)
@@ -41,7 +42,7 @@ parameters.set_matter_power(redshifts=zs, kmax=100.) # kmax should be much large
41
42
results = camb.get_results(parameters)
42
43
43
44
# HMcode
44
-
Pk = hmcode.power(k, zs, results)
45
+
Pk = hmcode.power(k, zs, results, T_AGN)
45
46
```
46
47
47
48
## Note
@@ -80,7 +81,7 @@ I think any residual differences between codes must therefore stem from:
80
81
But I didn't have time to investigate these differences more thoroughly. Note that there are accuracy parameters in `CAMB-HMcode` fixed at the $10^{-4}$ level, so you would never expect better than 0.01% agreement. Given that `HMcode` is only accurate at the ~2.5% level compared to simulated power spectra, the level of agreement between the codes seems okay to me, with the caveats above regarding very massive neutrinos.
81
82
82
83
While writing this code I had a few ideas for future improvements:
83
-
- Add the `HMcode-2020` baryon-feedback model; this would not be too hard for the enthusiastic student/postdoc.
84
+
-~~Add the `HMcode-2020` baryon-feedback model; this would not be too hard for the enthusiastic student/postdoc.~~ (Thanks Laila Linke!)
84
85
- The predictions are a bit sensitive to the smoothing $\sigma$ used for the dewiggling. This should probably be a fitted parameter.
85
86
- It's annoying having to calculate linear growth functions (all, LCDM), especially since the linear growth doesn't really exist. One should probably use the $P(k)$ amplitude evolution over time at some cleverly chosen scale instead, or instead the evolution of $\sigma(R)$ over time at some pertinent $R$ value. Note that the growth factors are *only* used to calculate the [Dolag et al. (2004)](https://arxiv.org/abs/astro-ph/0309771) correction and [Mead (2017)](https://arxiv.org/abs/1606.05345) $\delta_\mathrm{c}$, $\Delta_\mathrm{v}$.
86
87
- I never liked the halo bloating parameter, it's hard to understand the effect of modifying halo profiles in Fourier space. Someone should get rid of this (maybe modify the halo mass function instead?).
0 commit comments