Adding new parameters to CosmoMC, March-13

[alireza hojjati]

Hi Antony,

I am trying to add new parameters to CosmoMC. Since the structure has changed in this version and there is no definition of "num_hard" in the settings.f90 anymore, I tried to change it in cmbtypes.f90 as

num_hard = slow_num + 6

And add the new parameters right before the derived parameters (After #21: Aphiphi and #22:omegal). I get errors such as :

params_CMB.f90(188): error #6410: This name has not been declared as an array or a function. [PARAMS]
Params(22) = CMB%mag1
----- ^

Any suggestion?

Thanks,

Alireza

[Antony Lewis]

To add new theory parameters

* add the definition of the physical parameter to CMBParams in cmbtypes.f90

* Change params_CMB.f90 to add a new parameterization or modify an existing one (with a corresponding .paramnames file with the new parameter names in). In particular the _Init routine does

Code: Select all

call SetTheoryParameterNumbers(num_slow_params,num_semi_slow_params)

which sets the number of hard and semi-hard parameters (the latter being things like initial power spectrum parameters) and then

Code: Select all

  call this%Init(Ini,Names, 'params_CMB.paramnames')

to set the parameter names file.

* Modify CMB_Cls_Simple.f90 to pass your new CMBParams parameters to CAMB to actually change the result

* Edit your .ini files to to add range and width settings for the new parameter names


If instead your parameter is a likelihood parameter (nuisance parameter), the new parameter can be handled entirely inside your likelihood function code (see supernovae_SNLS.f90 for a simple example that has two nuisance parameters): you just set the .paramnames file (which determimes the number of new parameters), and then use them when passed in to your likelihood function.

[alireza hojjati]

Thanks, it worked !

[Tijmen de Haan]

My likelihood code includes likelihood/nuisance parameters, which I've successfully added as suggested.

My code is faster than the Planck likelihood, but slower than CAMB. Is it possible to assign these nuisance parameters their own fast/slow subspace?

[Antony Lewis]

Each likelihood object has a speed parameter, just set the appropriate number for the relative speed (CamSpec is set to have speed 5, higher number is faster).

[Tijmen de Haan]

Thanks. I noticed some code in paramdef.F90, and the clik_speed_* parameters in the clik-related ini files. How do I set the analogous speed parameter for my likelihood (which I added simply as an .ini file in batch1, a .paramnames file in data, a .f90 file in source, and an entry in DataLikelihoods.f90)?

[Antony Lewis]

In your f90 file defining your descendant of CosmologyLikelihood set the speed parateter. E.g.

Code: Select all

  module MyLlike
    use likelihood

    type, extends(CosmologyLikelihood) :: MyLikelihood
    ..
    contains
    ..
    end type MyLikelihood

..
    contains
 ..
    subroutine MyLike_Add(LikeList,Ini)
    class(LikelihoodList) :: LikeList
    Type(TIniFile) :: ini
    type(MyLikelihood), pointer :: like

     allocate(like)
     like%speed =  2
 ..

    call LikeList%Add(like)

...


end module

[Tijmen de Haan]

Awesome, thanks Antony!

[Andrei Lazanu]

Hi Antony,

Is the same true for the March 2014 version? I have modifeid the March 2013 according to your suggestions and it worked fine, but in the case of the March 2014 one the program compiles but does not seem to correctly read the parameter in the .ini file.

Best,
Andrei

[Antony Lewis]

A few things have been renamed. The template is now

Code: Select all

  module MyLlike
    use Likelihood_Cosmology

    type, extends(TCosmologyLikelihood) :: TMyLikelihood
    ..
    contains
    ..
    end type TMyLikelihood

..
    contains
 ..
    subroutine MyLike_Add(LikeList,Ini)
    class(TLikelihoodList) :: LikeList
    class(TSettingIni) :: ini
    type(TMyLikelihood), pointer :: like

     allocate(like)
     like%speed =  2
 ..

    call LikeList%Add(like)

...


end module

You can look at the various provided likelihoods for other examples.

[Andrei Lazanu]

Thanks Antony for the reply. In my case the parameter is a physical one so my question is if there is anything different compared to the previous versions apart from the obvious program name renamings, or if there is any other new place where I would have to add the parameter.

To add new theory parameters

* add the definition of the physical parameter to CMBParams in cmbtypes.f90

* Change params_CMB.f90 to add a new parameterization or modify an existing one (with a corresponding .paramnames file with the new parameter names in). In particular the _Init routine does
Code:
call SetTheoryParameterNumbers(num_slow_params,num_semi_slow_params)

which sets the number of hard and semi-hard parameters (the latter being things like initial power spectrum parameters) and then
Code:
call this%Init(Ini,Names, 'params_CMB.paramnames')

to set the parameter names file.

* Modify CMB_Cls_Simple.f90 to pass your new CMBParams parameters to CAMB to actually change the result

* Edit your .ini files to to add range and width settings for the new parameter names

[Antony Lewis]

The readme was out of date, I've now updated that section - thanks.

[Saisandri Saini]

Hii
I am trying to add "mnu" as a parameter to the chains and then use it in my Plots.How can I do it? I am using CosmoMC july 2019. Thanks

[Cristhian GQ]

Hello,

in the file batch3/params_CMB_defaults.ini, you need to change

param[mnu] = 0.06

for a fix value, to a a parameter by adding its center value, min value, max value, start width, and propose width. For example, replace the above by

param[mnu] = 0.06 0 0.9 0.005 0.005

or whatever values you want. Then CosmoMC will vary that parameter also.

Best regards,
Cristhian Garcia.

[Antony Lewis]

Or, better, set param[mnu] in the ini file for your run rather than changing the default one. Generally you shouldn't need to edit files in batch3, parameters can be overridden in the .ini file actually used.