Running CosmoMC without Planck likelihood

Use of Cobaya. camb, CLASS, cosmomc, compilers, etc.
Post Reply
Abir Sarkar
Posts: 16
Joined: July 07 2014
Affiliation: Raman Research Institute

Running CosmoMC without Planck likelihood

Post by Abir Sarkar » March 30 2016

Apart from Planck likelihood CosmoMC uses other likelihoods as BAODR11. WiggleZ, JLA etc.

Is it possible to run the code and plot results with just one of the likelihoods other than Planck? I have been able to run the chains and coverge but while running the ./getdist it can't read the chains.

Thanks,
Abir

Antony Lewis
Posts: 1941
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: Running CosmoMC without Planck likelihood

Post by Antony Lewis » March 31 2016

You'll need to be more specific about what exactly you are running and what exactly goes wrong.

Abir Sarkar
Posts: 16
Joined: July 07 2014
Affiliation: Raman Research Institute

Running CosmoMC without Planck likelihood

Post by Abir Sarkar » March 31 2016

I am running the code with the WiggleZ likelihood only. I have disabled the Planck lowlike, lowl and CAMSpec likelihoods. I am taking runs with standard parameters only and did not make any modification in the source codes.

While I am running ./cosmomc with .ini file it's running fine and converging.

Then while trying to run the ./getdist then it's showing the following error

" error reading line 0 - skipping to next row"

which is usually seen when the column number in the file_root.txt files does not match with the number provided in the distparams.ini file. But I have specified the columnnum to be 0 so there should not be a problem.

This problem does not occur when I include Planck likelihoods with any other likelihoods.

One thing I have noticed that the software uses a different sampling method in this case. Should there be any modification regarding that?

Antony Lewis
Posts: 1941
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: Running CosmoMC without Planck likelihood

Post by Antony Lewis » March 31 2016

I would look at the chain file in a text file to see what's going wrong, (e.g. NaN or something else).

Post Reply