bug with neutrino parameters - CAMB online works fine, downloaded CAMB produces "forrtl: severe (174): SIGSEGV, segmentation fault occurred"

[Martina Schwind]

Dear Antony and friends,

12.06. 11:35:
even when I set
massive_neutrinos = 1
it doesn't work...



12.06. - 10:35:
the reason must be the following parameter:
nu_mass_fractions = 0 1

with the following parameter it works:
nu_mass_fractions = 0.0000001 0.9999999

with this it doesn't work as well:
nu_mass_fractions = 0.0 1.0


12.06. - 10:20:

I test the following parameters:
output_root = /server/lambda/htdocs/camb/output/camb_17473176
number_of_threads = 0
get_scalar_cls = T
get_vector_cls = F
get_tensor_cls = F
get_transfer = F
do_lensing = T
do_nonlinear = 0
simul_type = 0
nsmax = 512
fwhm_arcmin = 420
iseed = 0
l_max_scalar = 2000
k_eta_max_scalar = 4000
l_max_tensor = 1500
k_eta_max_tensor = 3000
use_physical = T
hubble = 70
temp_cmb = 2.725
ombh2 = 0.0226
omch2 = 0.114
omnuh2 = 0.00132
omk = 0
nu_mass_eigenstates = 2
nu_mass_degeneracies = 2 1
nu_mass_fractions = 0 1
omega_baryon = 0.046
omega_cdm = 0.224
omega_lambda = 0.73
omega_neutrino = 0
helium_fraction = 0.24
massless_neutrinos = 0
massive_neutrinos = 3
w = -1
cs2_lam = 1
reionization = T
re_use_optical_depth = F
re_optical_depth = 0.09
re_redshift = 11
re_delta_redshift = 0.5
re_ionization_frac = 1
initial_power_num = 1
scalar_amp(1) = 2.46e-9
scalar_spectral_index(1) = 0.96
scalar_nrun(1) = 0
tensor_spectral_index(1) = 0
initial_ratio(1) = 1
initial_condition = 1
initial_vector = -1 0 0 0 0
vector_mode = 0
COBE_normalize = F
CMB_outputscale = 7.43e12
pivot_scalar = 0.002
pivot_tensor = 0.002
transfer_high_precision = F
transfer_interp_matterpower = T
transfer_kmax = 2
transfer_k_per_logint = 5
transfer_num_redshifts = 1
transfer_redshift(1) = 0
feedback_level = 1
lensing_method = 1
accurate_BB = F
accurate_reionization = F
do_tensor_neutrinos = F
massive_nu_approx = 0
accurate_polarization = T
do_late_rad_truncation = T
RECFAST_fudge = 1.14
RECFAST_fudge_He = 0.86
RECFAST_Heswitch = 6
accuracy_boost = 1
l_accuracy_boost = 1
l_sample_boost = 1
scalar_output_file = scalcls.dat
vector_output_file = vectcls.dat
tensor_output_file = tenscls.dat
total_output_file = totcls.dat
lensed_output_file = lensedcls.dat
lensed_total_output_file = lensedtotcls.dat
FITS_filename = scalcls.fits
transfer_filename(1) = transfer_out_z0.dat
transfer_matterpower(1) = matterpower_z0.dat



via CAMB online - which works fine:

=============== CAMB Processing ===============

Age of universe/GYr = 13.703
Om_b h^2 = 0.022600
Om_c h^2 = 0.114000
Om_nu h^2 = 0.001320
Om_Lambda = 0.718531
Om_K = 0.000000
Om_m (1-Om_K-Om_L) = 0.281469
100 theta (CosmoMC) = 1.045333
2.00 nu, m_nu*c^2/k_B/T_nu0 = 0.00 (m_nu = 0.000 eV)
1.00 nu, m_nu*c^2/k_B/T_nu0 = 741.21 (m_nu = 0.125 eV)
tau_recomb/Mpc = 284.33 tau_now/Mpc = 14254.3


and via CAMB downloaded - where I get the following error:

Om_b h^2 = 0.022600
Om_c h^2 = 0.114000
Om_nu h^2 = 0.001320
Om_Lambda = 0.718531
Om_K = 0.000000
Om_m (1-Om_K-Om_L) = 0.281469
100 theta (CosmoMC) = 1.045333
2.0000 nu, m_nu*c^2/k_B/T_nu0 = 0.00 (m_nu = 0.000 eV)
1.0000 nu, m_nu*c^2/k_B/T_nu0 = 741.21 (m_nu = 0.124 eV)
Age of universe/GYr = 13.703
zstar = 1089.06
r_s(zstar)/Mpc = 146.06
100*theta = 1.045505
zdrag = 1060.09
r_s(zdrag)/Mpc = 148.67
k_D(zstar) Mpc = 0.1398
100*theta_D = 0.160902
z_EQ (if v_nu=1) = 3286.74
100*theta_EQ = 0.838837
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
camb 000000000060003F Unknown Unknown Unknown
camb 0000000000603F2E Unknown Unknown Unknown
camb 0000000000602A82 Unknown Unknown Unknown
camb 00000000005A4221 Unknown Unknown Unknown
camb 00000000005A1B91 Unknown Unknown Unknown
camb 0000000000594727 Unknown Unknown Unknown
camb 000000000040043C Unknown Unknown Unknown
camb 000000000052583E Unknown Unknown Unknown
camb 0000000000400309 Unknown Unknown Unknown
schwind@pi208f:~/master-project/camb$

So it didn't calculate the following:

tau_recomb/Mpc = 284.03 tau_now/Mpc = 14159.0
tau_recomb/Mpc = 283.95 tau_now/Mpc = 14159.0
at z = 0.0000000E+00 sigma8 (all matter)= 0.7400333


Do You have any idea what the reason could be for this error?

regards martina

[Antony Lewis]

I would use two massless neutrinos and one massive eigenstate for that case. (I guess it should automatically map your parameters into that, currently the crash is because it assumes eigenstate masses are positive and non-zero, i.e. divides by them in some places to estimate time when they become non-relativistic).

[Martina Schwind]

I used no the following parameters:

nu_mass_eigenstates = 1
nu_mass_degeneracies = 1
nu_mass_fractions = 1

massless_neutrinos = 2
massive_neutrinos = 1

and this works.



thanks Martina