agnwinds / python

This is the repository for Python, the radiative transfer code used to winds in AGN and other syatems
GNU General Public License v3.0
26 stars 25 forks source link

SV model Lyman edge slight regression disagreement #46

Closed jhmatthews closed 11 years ago

jhmatthews commented 11 years ago

I'm just regression testing a release and wanted to flag something up as potentiallyproblematic before I release it. I was plotting a CV model, and comparing it to 74b4, which was when I carried out a pretty rigorous regression test. 76b is in green and 74b4 in blue.

sv76b_compsources

sv76b_compsummary

The thing I was concerned about was the difference beyond the lyman edge and if we understand why there is a difference here- I couldn't spot anything specific in what ahs changed over the versions

Should I be worrying about this? Or can I go ahead and make a 76b release which will then be used as our testing baseline?

These were run with identical pf files and the only difference was that sv76b was run in parallel. I will do a non parallel run to check if that is the problem.

jhmatthews commented 11 years ago

pf file:

Wind_type(0=SV,1=Sphere,2=Previous,3=Proga,4=Corona,5=knigge,6=thierry,7=yso,8=elvis,9=shell)   0
Atomic_data atomic/standard39
photons_per_cycle   100000
Ionization_cycles   20
spectrum_cycles 10
Coord.system(0=spherical,1=cylindrical,2=spherical_polar,3=cyl_var) 1
Wind.dim.in.x_or_r.direction    30
Wind.dim.in.z_or_theta.direction    30
Wind_ionization(0=on.the.spot,1=LTE,2=fixed,3=recalc_bb,5=recalc_pow,6=pairwise_bb,7=pairwise_pow)  3
Line_transfer(0=pure.abs,1=pure.scat,2=sing.scat,3=escape.prob,6=macro_atoms,7=macro_atoms+aniso.scattering)    3
System_type(0=star+disk,1=star+disk+secondary,2=agn)    0
Star_radiation(y=1) 1
Disk_radiation(y=1) 1
Boundary_layer_radiation(y=1)   0
Wind_radiation(y=1) 1
Rad_type_for_star(0=bb,1=models)_to_make_wind   0
Rad_type_for_disk(0=bb,1=models)_to_make_wind   0
Rad_type_for_bl(0=bb,1=models)_to_make_wind 0
Overall_system_type(0=single_object,1=binary)   0
mstar(msol) 0.8
rstar(cm)   7e+08
tstar   40000
disk.type(0=no.disk,1=standard.flat.disk,2=vertically.extended.disk)    1
disk.mdot(msol/yr)  1e-08
Disk.illumination.treatment(0=no.rerad,1=high.albedo,2=thermalized.rerad,3=analytic)    0
Disk.temperature.profile(0=standard;1=readin)   0
disk.radmax(cm) 2.4e+10
lum_bl  0
t_bl    10000
Torus(0=no,1=yes)   0
wind.radmax(cm) 1e+12
wind.t.init 40000
wind.mdot(msol/yr)  1e-09
sv.diskmin(wd_rad)  4
sv.diskmax(wd_rad)  12
sv.thetamin(deg)    20
sv.thetamax(deg)    65
sv.mdot_r_exponent  0
sv.v_infinity(in_units_of_vescape   3
sv.acceleration_length(cm)  7e+10
sv.acceleration_exponent    1.5
Rad_type_for_star()_in_final_spectrum   0
Rad_type_for_disk(0=bb,1=models,2=uniform)_in_final_spectrum    0
spectrum_wavemin    850
spectrum_wavemax    1850
no_observers    5
angle(0=pole)   10
angle(0=pole)   27.5
angle(0=pole)   45
angle(0=pole)   62.5
angle(0=pole)   80
phase(0=inferior_conjunction)   0.5
phase(0=inferior_conjunction)   0.5
phase(0=inferior_conjunction)   0.5
phase(0=inferior_conjunction)   0.5
phase(0=inferior_conjunction)   0.5
live.or.die(0).or.extract(anything_else)    1
Select_specific_no_of_scatters_in_spectra(y/n)  n
Select_photons_by_position(y/n) n
spec.type(flambda(1),fnu(2),basic(other)    1
Extra.diagnostics(0=no) 0
Use.standard.care.factors(1=yes)    1
Photon.sampling.approach(0=T,1=(f1,f2),2=cv,3=yso,4=user_defined)   2
kslong commented 11 years ago

I assume you mean the relatively small difference below 912 A in the continuum. This indicates a slight change in the amount of neutral H. I've seen similar things in the past, and indeed this might be associated with the free-free fix for which you saw less emission. I am not concerned about this.

Knox

Knox S. Long Space Telescope Science Institute

jhmatthews commented 11 years ago

Ok, great, issue closed and release imminent.