CAMB, "forrtl: severe (174): SIGSEGV, segmentation fault occurred", when the background evolution is changed

Use of Healpix, camb, CLASS, cosmomc, compilers, etc.
Post Reply
ehsan sadri
Posts: 7
Joined: August 19 2018
Affiliation: IAUCTB

CAMB, "forrtl: severe (174): SIGSEGV, segmentation fault occurred", when the background evolution is changed

Post by ehsan sadri » September 09 2018

I would like to change the background evolution. there are many different background models rather than the standard cosmology.
I have examined all of them but CAMB send an error
forrtl: severe (174): SIGSEGV, segmentation fault occurred
I tested the CAMB with the original code. I just multiplied one "a" in "dtauda. in CAMB we have

Code: Select all

dtauda=sqrt(3/grhoa2)
I changed it to

Code: Select all

dtauda=sqrt(3/(a*grhoa2))
but even for this term CAMB sent the same error. or even for

Code: Select all

dtauda=2/(0.846+sqrt(grhoa2))

So, I really do not understand what should I do to change the background.
the full error is:
1 nu, g= 1.0153 m_nu*c^2/k_B/T_nu0= 353.71 (m_nu= 0.060 eV)
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
camb 00000000005617E1 Unknown Unknown Unknown
camb 000000000055FF37 Unknown Unknown Unknown
camb 0000000000466814 Unknown Unknown Unknown
camb 0000000000466626 Unknown Unknown Unknown
camb 0000000000403C04 Unknown Unknown Unknown
camb 0000000000407DCD Unknown Unknown Unknown
I appreciate any help.
Best,

Antony Lewis
Posts: 1387
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: CAMB, "forrtl: severe (174): SIGSEGV, segmentation fault occurred", when the background evolution is changed

Post by Antony Lewis » September 10 2018

You can compile with "make Debug" to potentially get a more useful error message. Note that the background is evolved both using the dtauda function, as in the derivs functions in equation(_pff).f90, so they need to be consistent.

ehsan sadri
Posts: 7
Joined: August 19 2018
Affiliation: IAUCTB

Re: CAMB, "forrtl: severe (174): SIGSEGV, segmentation fault occurred", when the background evolution is changed

Post by ehsan sadri » September 10 2018

I will do it to see the more accurate error, and post here the result. But as you mentioned, the changes in background should be identical in both equations.f90 and equations_ppf.f90 files, it is potential to solve these kind of problems?

ehsan sadri
Posts: 7
Joined: August 19 2018
Affiliation: IAUCTB

Re: CAMB, "forrtl: severe (174): SIGSEGV, segmentation fault occurred", when the background evolution is changed

Post by ehsan sadri » September 12 2018

Antony Lewis wrote:
September 10 2018
You can compile with "make Debug" to potentially get a more useful error message.
I did make Debug. It prints a long list of useless errors. But at the end, it prints:
/tmp/ifortMQwBiU.i90(342): catastrophic error: Too many errors, exiting
compilation aborted for inidriver.F90 (code 1)
Makefile_main:79: recipe for target 'camb' failed
make: *** [camb] Error 1


I just multiplied one [math] into the function of [math] in equations_ppf.f90 file. Really, isn't it the right way of changing the background evolution? there is no other way so.
Thank you for you help, in advance.


Post Reply