Strange behavior in modules.f90 with function dtauda (cosmomc package)

Use of Cobaya. camb, CLASS, cosmomc, compilers, etc.
Post Reply
Patrice Okouma
Posts: 19
Joined: November 05 2009
Affiliation: University of Cape Town

Strange behavior in modules.f90 with function dtauda (cosmom

Post by Patrice Okouma » December 18 2009

Hi all --- This is my problem with cosmomc. We recently included 1 extra parameter. cosmomc compiles fine.

When running it, the first call for TimeOfz(z) is done with z = 0._dl : CP%tau0=TimeOfz(0._dl) , line 342 in modules.f90 . It leads to the computation of DeltaTime=rombint(dtauda,a1,a2,atol) where a1 = 0._dl and a2 = 1._dl.
It seems that in its first call, dtauda(a) has problem when computing for a = 1._dl .
dtauda(1._dl) gives nothing out .
THEN I did a dummy test by having it ran at ( 1._dl - 0.0001 ) .... The new error message is now :
Likelihood: 1.0000000E+30 Current Like: 1.0000000E+30
1001 rejecting. ratio: 0.0000000E+00
MCMC.f90: Couldn't start after 1000 tries - check starting ranges

Could there be a simple reason why dtauda(a) has problem when computing for a = 1._dl

Thanks
Patrice

Post Reply