Page 1 of 1

[CosmoMC] - changed n_run convention?

Posted: August 05 2017
by Ira *Wolfson

My old rig has fallen, so I just finished basically formatting everything and reinstalling CosmoMC (and all deps).

I have now run 3 separate runs with the different datasets.

Just as a reference:

Code: Select all

#general settings
#Bicep-Keck-Planck, varying cosmological parameters

#Planck 2015, default just include native likelihoods (others require clik)

#Other Likelihoods

#general settings

#e.g. to vary r in addition to standard 6:
#(for r>0 also need compute_tensors=T)
compute_tensors = T
param[r] = 0.06 0.0 0.5 0.04 0.04
param[logA] = 3.1 1.5 4.5 0.04 0.04
param[ns] = 0.96 0.9 1.05 0.04 0.04
param[nrun]=-0.09 -0.5 0.5 0.04 0.04
#limits[r02]=0 N
limits[r]=0 N

I used different data sets - including BAO, Wigglez etc.

The results for [tex]n_{run}[/tex] are always slightly positive, where in my previous experience, when [tex]n_{run}[/tex] is used (and [tex]n_{run,run}[/tex] is set to 0 ) , one usually gets a slightly negative value.

This is also the case with the most recent Planck collaboration analysis.

Has the convention for CosmoMC changed between the last stable version and the current master branch?

If not, any advise on what is happening?


-- Ira

Re: [CosmoMC] - changed n_run convention?

Posted: August 06 2017
by Antony Lewis
As far as I know nothing has changed (could check by compare outputs for non-trivial initial nrun using action=4 for last stable versus master).