COSMOMC: Changing input parameter

Use of Healpix, camb, CLASS, cosmomc, compilers, etc.
Post Reply
Jostein R. Kristiansen
Posts: 4
Joined: February 14 2006
Affiliation: University of Oslo
Contact:

COSMOMC: Changing input parameter

Post by Jostein R. Kristiansen » February 14 2006

Hi,

I'm using cosmomc, and would like to use omnuh2 as an input parameter instead of nufrac. I've changed from CMB%nufrac = Params(6) to CMB%omnuh2 = Params(6) in the params_CMB.f90 file and changed the limits in my params.ini file accordingly. When I'm running the chains, the limits on omnuh2 become a lot larger (~ 0.5) than one should expect, although the limits I got when using nufrac as an input parameter seemed perfectly reasonable. What am I doing wrong?

Thanks. Jostein.

Anze Slosar
Posts: 183
Joined: September 24 2004
Affiliation: Brookhaven National Laboratory
Contact:

COSMOMC: Changing input parameter

Post by Anze Slosar » February 14 2006

I think you need to change the reverse mapping as well, i.e in cmbparastoparams, put something like
Params(6) =CMB%omnuh2

Jostein R. Kristiansen
Posts: 4
Joined: February 14 2006
Affiliation: University of Oslo
Contact:

COSMOMC: Changing input parameter

Post by Jostein R. Kristiansen » February 14 2006

I forgot to mention it, but I've changed that part as well. And still the results look strange. Jostein.

Antony Lewis
Posts: 1522
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: COSMOMC: Changing input parameter

Post by Antony Lewis » February 14 2006

Also make sure you set up CMB%nufrac from the input CMB%omnuh2, or edit the later lines where it does

Code: Select all

    CMB%omnuh2 = CMB%omdmh2*CMB%nufrac
    CMB%omch2 = CMB%omdmh2 - CMB%omnuh2

Jostein R. Kristiansen
Posts: 4
Joined: February 14 2006
Affiliation: University of Oslo
Contact:

COSMOMC: Changing input parameter

Post by Jostein R. Kristiansen » February 23 2006

Thank you. Sounds of course reasonable. I changed

Code: Select all

 CMB%omnuh2 = CMB%omdmh2*CMB%nufrac 
to

Code: Select all

 CMB%nufrac = CMB%omnuh2/CMB%omdmh2 
and now the results look good. Thanks a lot. Jostein.

Post Reply