SPheno error with two-scale matching (flag 66) and micrOMEGAs output

[Socob]

Looking at the singlet–doublet fermion dark matter model (e. g. arXiv:1109.2604), I have the problem that when both

• two-scale matching (flag 66 = 1 in the LesHouches input file) and
• output for micrOMEGAs (flag 50 = 0 and/or flag 77 = 1)

are activated, I get the following SPheno error for certain parameter values:

Code: Select all

 Calculating mass spectrum
              1 .-iteration
Note: The following floating-point exceptions are signalling: IEEE_INVALID_FLAG IEEE_DIVIDE_BY_ZERO IEEE_UNDERFLOW_FLAG IEEE_DENORMAL
STOP Subroutine TerminateProgram

If either two-scale matching or micrOMEGAs output is disabled, this error does not appear. I have attached my SARAH model files for reference.

In particular, the error appears for mdoubInput > 0, but not for mdoubInput < 0. For example:

Code: Select all

Block MINPAR      # Input parameters 
 1   2.7000000E-01    # lambdaInput
 12   6.1000000E+02    # msingInput
 13   7.2000000E+02    # mdoubInput
 21   1.1000000E-02    # lambda1Input
 22   1.2000000E-02    # lambda2Input

What could be the problem with positive values for that particular parameter?

Version information: SARAH 4.12.3, SPheno 4.0.3.

[FStaub]

Hi,

the problem was that it tried to take the log of a negative mass. I have added the missing abs(). The patch is attached.

In addition, I think you should also change your model and add a phase for your vector-like states:

Code: Select all

DEFINITION[EWSB][Phases]= 
{    {doub1p, PhaseD}
    }; 

Otherwise, you'll see that the results between negative and positive mass conventions won't agree.

Cheers,
Florian

[BjHerrmann]

Hi Florian,
I was about to ask a similar question when I saw this post...
I have the following issue when running SPhenoMSSM generated using model MSSM in SARAH 4.12.3 and SPheno 4.0.3

~/Projects/GUT_NMFV/SU5xA4> ./SPhenoMSSM test.in
Warning from routine CalculateMSu
in the calculation of the masses
occurred a negative mass squared!
1 -283723.72038028348
Calculating mass spectrum
1 .-iteration
NaN appearing in CouplingcUFdGluSd
Note: The following floating-point exceptions are signalling: IEEE_INVALID_FLAG IEEE_DIVIDE_BY_ZERO IEEE_UNDERFLOW_FLAG IEEE_DENORMAL
STOP Subroutine TerminateProgram

In addition, there is also something about a NaN in some coupling...
Could this be the same issue?

I attach my test.in for information.

Cheers from Annecy,
Björn

[FStaub]

Hi Bjoern,

no, I think your problem is another one. It seems that you are using an input which was originally made for SPheno? SPhenoMSSM might need some more information in the block SPhenoInput. Also, you need to define the entire matrices for Sfermion soft masses (it doesn't add the symmetric pieces by it's own). Note, SARAH will write template files which show you what is needed.

Cheers,
Florian

[BjHerrmann]

Hi Florian,
thanks for your reply! Then it comes probably from the fact that I did not specify the full 6x6 soft mass matrices. I will check a.s.a.p.
Cheers,
Björn

[Socob]

Hi,

the problem was that it tried to take the log of a negative mass. I have added the missing abs(). The patch is attached.

In addition, I think you should also change your model and add a phase for your vector-like states:

Code: Select all

DEFINITION[EWSB][Phases]= 
{    {doub1p, PhaseD}
    }; 

Otherwise, you'll see that the results between negative and positive mass conventions won't agree.

Cheers,
Florian

Thanks for the quick fix, and also for the hint about the phase!

Unfortunately, it seems that I’m getting a different error now. With output for micrOMEGAs (flag 50 = 0, flag 77 = 1), I see

Code: Select all

$ bin/SPhenoTest input/LesHouches.in.Test_low
 Calculating mass spectrum
              1 .-iteration
   ... reached precision:   1.0000000000000000     
              2 .-iteration
   ... reached precision:   1.0959174872941060E-005
 Calculating branching ratios and decay widths
 Calculating one loop decays
 Loop masses not calculated: tree-level masses used for kinematics
 Loop masses not calculated: no U-factors are applied
 Calculating one-loop decays of Fu 
 Calculating one-loop decays of Fe 
 Calculating one-loop decays of Fd 
 Calculating one-loop decays of hh 
 Calculating one-loop decays of Hp 
 Calculating one-loop decays of Fneu 
 Please have a look if the decay you are considering is kinematically
 allowed. Error message by Hard photon emission!
 Please have a look if the decay you are considering is kinematically
 allowed. Error message by Hard photon emission!
 Please have a look if the decay you are considering is kinematically
 allowed. Error message by Hard photon emission!
 Calculating one-loop decays of Fcha 
 Calculating low energy constraints
 Writing output files
 Finished!

However, when micrOMEGAs output is disabled (flag 50 = 1, flag 77 = 0), I get

Code: Select all

$ bin/SPhenoTest input/LesHouches.in.Test_low
 Calculating mass spectrum
              1 .-iteration
   ... reached precision:   1.0000000000000000     
              2 .-iteration
   ... reached precision:   1.0959174872941060E-005
 Calculating branching ratios and decay widths
 Calculating one loop decays
 Loop masses not calculated: tree-level masses used for kinematics
 Loop masses not calculated: no U-factors are applied
 Calculating one-loop decays of Fu 
 Calculating one-loop decays of Fe 
 Calculating one-loop decays of Fd 
 Calculating one-loop decays of hh 
 Calculating one-loop decays of Hp 
 Calculating one-loop decays of Fneu 
 Calculating one-loop decays of Fcha 
 Calculating low energy constraints
 Writing output files
 Finished!

[FStaub]

Hi,

it seems that this is another case where the possibility of negative masses wasn't thought of, i.e. Abs() missing again. I don't have time right now to run it my own, but I guess the patch in the attachment should work... if not, please come back.

Cheers,
Florian

[Socob]

Hi,

it seems that this is another case where the possibility of negative masses wasn't thought of, i.e. Abs() missing again. I don't have time right now to run it my own, but I guess the patch in the attachment should work... if not, please come back.

Cheers,
Florian

It seems to work now – thanks again!