Dear all,
I have added three new slow parameters following the information in the Readme file of the COSMOMC and in several similar topics here. I summarize the changes I have made at the end of this message. I tried running COSMOMC with both options; asking for estimating a new covariance matrix and also with the old matrix filled with zeros for the added entries. In any case I got the following output on the screen (this is an example using 6 processors)
===
Number of MPI processes: 6
Random seeds: 2354, 15909 rand_inst: 1
Random seeds: 12426, 30014 rand_inst: 5
Random seeds: 30414, 24716 rand_inst: 3
Computing tensors: F
Doing CMB lensing: T
lmax = 2100
Number of C_ls = 3
Random seeds: 17023, 772 rand_inst: 6
Random seeds: 5732, 22170 rand_inst: 4
Random seeds: 8613, 796 rand_inst: 2
Computing tensors: F
Doing CMB lensing: T
lmax = 2100
Number of C_ls = 3
Computing tensors: F
Doing CMB lensing: T
lmax = 2100
Number of C_ls = 3
Computing tensors: F
Doing CMB lensing: T
lmax = 2100
Number of C_ls = 3
Computing tensors: F
Doing CMB lensing: T
lmax = 2100
Number of C_ls = 3
Computing tensors: F
Doing CMB lensing: T
lmax = 2100
Number of C_ls = 3
===
and nothing else, even after a couple of days. I have checked the chains and all of they are empty as well as the log files.
I would appreciate any help.
Best,
César.
====
Summary of modifications to COSMOMC
====
The new parameters are defined in modules.f90 of CAMB as
type CAMBparams
real(dl) :: qVpar1, qVpar2, qVpar3

In COSMOMC they are added with
In settings.f90
integer, parameter :: num_hard =10

In cmbtypes.f90
Type CMBParams
real Q1, Q2, Q3

In params_CMB.f90
subroutine SetForH(Params,CMB,H0)
CMB%Q1 = Params(8)
CMB%Q2 = Params(9)
CMB%Q3 = Params(10)
subroutine CMBParamsToParams(CMB, Params)
Params(8) = CMB%Q1
Params(9) = CMB%Q2
Params(10) = CMB%Q3

In CMB_Cls_simple.f90
subroutine CMBToCAMB(CMB,P)
P%qVpar1 = CMB%Q1
P%qVpar2 = CMB%Q2
P%qVpar3 = CMB%Q3

In params.ini
param8 = 10 100 1000 10 10
param9 = 0.01 0.001 0.1 0.01 0.01
param10 = 0 0 0 0 0
#n_s
param11 = 0.95 0.5 1.5 0.02 0.01
#n_t
param12 = 0 0 0 0 0
#n_run
param13 = 0 0 0 0 0
#log[10^10 A_s]
param14 = 3 2.7 4 0.01 0.01
#amp_ratio
param15 = 0 0 0 0 0
#SZ amplitude, as in WMAP analysis
param16 = 1 0 2 0.4 0.4
adding parameters, empty chains

 Posts: 5
 Joined: April 16 2008
 Affiliation: CINVESTAV
adding parameters, empty chains
I don't know that this is the only issue, but this looks like a problem:
The starting value isn't in the range provided.param8 = 10 100 1000 10 10

 Posts: 5
 Joined: April 16 2008
 Affiliation: CINVESTAV
adding parameters, empty chains
Dear Ben,
Thank you. Indeed, that silly error was the source of the problem.
Best regards,
César.
Thank you. Indeed, that silly error was the source of the problem.
Best regards,
César.