A strange likelihood value

Use of Cobaya. camb, CLASS, cosmomc, compilers, etc.
Post Reply
bin hu
Posts: 23
Joined: September 18 2008
Affiliation: leiden univ

A strange likelihood value

Post by bin hu » July 20 2015

Hi Antony,

I simply install plc2.0 and run cosmomc_july15 with the following input:
The .log file shows: CAMB called 100 times; 0 errors

More strange to me, it shows a strange likelihood value, like:

Dragging time: 1.25476098060608
drag steps accept ratio: 0.2558140
Chain:2 drag accpt: 0.3916084 fast/slow 64.89683 slow: 126
3 Fast dragging, Like: 5654.89136515482
Calling CAMB
CAMB done
lowl_SMW_70_dx11d_2014_10_03_v5c_Ap lnlike = 5247.73176408751
DR11LOWZ BAO likelihood = 1.23941947323490
commander_rc2_v1.1_l2_29_B lnlike = 6.41255429423666
6DF BAO likelihood = 0.164469178994089
plik_dx11dr2_HM_v18_TT lnlike = 394.848751938114
Dragging Slow time: 1.75624680519104
plik_dx11dr2_HM_v18_TT lnlike = 399.393982273100
plik_dx11dr2_HM_v18_TT lnlike = 397.519798566244

If I understand well, plik_dx11dr2_HM_v18_TT lnlike should be around thousands, and lowl should be much smaller.

Does this means my plc2.0 is not installed properly?

Many thanks!
Bin

---------------------------
.INI
#general settings
#Bicep-Keck-Planck, varying cosmological parameters
#DEFAULT(batch2/BKPlanck.ini)

#Planck 2015, default just include native likelihoods (others require clik)
DEFAULT(batch2/plik_dx11dr2_HM_v18_TT.ini)
DEFAULT(batch2/lowTEB.ini)
DEFAULT(batch2/lowl.ini)
DEFAULT(batch2/lensing.ini)

#Other Likelihoods
DEFAULT(batch2/BAO.ini)
#DEFAULT(batch2/WiggleZ_MPK.ini)
#DEFAULT(batch2/MPK.ini)
#DEFAULT(batch2/WL.ini)

#general settings
DEFAULT(batch2/common.ini)

#e.g. to vary r in addition to standard 6:
#(for r>0 also need compute_tensors=T)
#compute_tensors = T
#param[r] = 0.03 0 2 0.04 0.04

#high for new runs
MPI_Max_R_ProposeUpdate = 30

propose_matrix= planck_covmats/base_TT_lowTEB_plik.covmat

#Folder where files (chains, checkpoints, etc.) are stored
root_dir = chains/

#Root name for files produced
file_root=plik_lowteb_lowl_lensing_bao
#action= 0 runs chains, 1 importance samples, 2 minimizes
#use action=4 just to quickly test likelihoods
action = 0

num_threads = 0

#if you want to get theory cl for test point
#test_output_root = output_cl_root

start_at_bestfit =F
feedback=2
use_fast_slow = T

checkpoint = F

#sampling_method=7 is a new fast-slow scheme good for Planck
sampling_method = 7
dragging_steps = 3
propose_scale = 2
---------------------------

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

Re: A strange likelihood value

Post by Antony Lewis » July 20 2015

You can run ./cosmomc test_planck.ini to check that likelihoods are running as expected.

The default high-L 2015 likelihood is binned, so the chi2 is much lower than the unbinned 2013 likelihood.

The lowl_SMW is a pixel likelihood, so a large chi2. Note that you don't want both lowTEB (which is T+E+B at low L) and lowl (which is just T at low L).

bin hu
Posts: 23
Joined: September 18 2008
Affiliation: leiden univ

A strange likelihood value

Post by bin hu » August 06 2015

Hi Antony,

Are there available script for using unbinned plik?

Many thanks!
Bin

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

Re: A strange likelihood value

Post by Antony Lewis » August 07 2015

No, but should be simple to adapt the binned one.

bin hu
Posts: 23
Joined: September 18 2008
Affiliation: leiden univ

A strange likelihood value

Post by bin hu » August 07 2015

Which value should I set for 'clik_speed_plik' if I want to use "plik_dx11dr2_HM_v18_TT_bin1"?

Thanks!

bin hu
Posts: 23
Joined: September 18 2008
Affiliation: leiden univ

A strange likelihood value

Post by bin hu » August 07 2015

Hi Antony,

Does this likelihood value sound reasonable with _bin1?

adding parameters for: plik_dx11dr2_HM_v18_TT_bin1
Fast divided into 1 blocks
21 parameters ( 6 slow ( 0 semi-slow), 15 fast ( 0 semi-fast))
Time for theory: 4.98201
Time for plik_dx11dr2_HM_v18_TT_bin1: 0.341078996658325
loglike chi-sq
4402.514 8805.028 CMB: plik = plik_dx11dr2_HM_v18_TT_bin1
Test likelihoods done, total logLike, chi-eq = 4402.639 8805.278
Likelihood calculation time (seconds)= 5.4304
Total time: 8 ( 0.00211 hours )

Cheers, Bin

Post Reply