Closed vrastil closed 6 years ago
fixing in branch fix_mesh_pwr
now working for Zel'dovich
change of
m = pow(Ng, 3)
to
m = pow(Ng, 1.5)*pow(Ng_pwr, 1.5)
Why? need to investigate exact units and \rho_avg = 1
Done
SOLUTION:
m = pow(Ng_pwr, 3)
- justified from overdensity definiton, no need for \rho_avg = 1
Extra normalization of power spectra Pk *= pow(Ng, 3)
- probably due to FFT definiton and FT of gaussian random field (n(x, \sigma = 1) -> sqrt(2*N) n(k, \sigma = 1)
power_aux
as temporary store of potentialSolve with resizing of Mesh
? Use std::vector
instead of new
?
WON`T FIX: In order to have proper resolution one should have inequalities
N_a >= N_p >= N_p
In order to have compatible results for various simulations, the power spectrum evaluation should be performed on the same grid. N_m = 1024