Hi Florian, I think I found a potential bug in SARAH.

The model I am working on contains a Yukawa parameter (called 'Yep') that carries three indices. I traced the problems back to this parameter, which seems to lead to inconsistencies in line 283 of nonSUSYrges.m. In particular, SARAH fails when it tries getDimParameter[threeIndexParameter[]] because Position[parameters,x] returns nothing. This is because the Yukawa parameter 'Yep' appears as 'Yep' in the parameters array, while threeIndexParameter[] reads 'Yep(gt2, gt1,gt3)'. This mismatch seems to be at the origin of the problem.

I tried replacing
in line 276 of nonSUSYrges.m by
but now I get a different error:
And therefore this solution is not satisfactory.

I attach the model files in case you find them necessary for testing.

By the way, the model also contains 4-index parameters and a 1-index VEV. This could perhaps lead to problems as well, as these are unconventional features. Of course, a solution would be to split the 3-generation scalars into 3 independent fields, but that's of course not optimal.

Well, you need to add the parameters, see e.g MSSM-RpV/TriRpV/parameters.m at the very end.

Cheers,
Florian

Hmmm, a parameter with four indices is so far not supported. In that case it might work to run
Then you also don't need to define the additional parameters coming from the splitting of tensors into a set of matrices.

Not sure I understood the solution, sorry.

Since I also have 4-index parameters: just properly defining the parameters in the parameters.m file (as for trilinear RPV) would not be sufficient. Therefore, you propose

1) To change the nonSUSYrges.m file as explained above
2) To run MakeSPheno[] after CalcRGEs[NoMatrixMultiplication -> True]

is this correct? Because I tried this way and now I get error messages in CalcRGEs.

Thanks in advance.

Hi Avelino,

checking your model, there is another problem beside the ones with quartics with four indices:

The RGEs are calculated for real scalars, ie. each complex scalar is decomposed by SARAH. That means, when it considers the diagrams with external leges re(H), re(H), re(Phi), re(Phi) the amplitude from that diagram contributes to the running of lambda5,6,7. Therefore, to single out the contribution to the running of one coupling, it must calculate the superposition of different combinations of re(X) and im(X). To find the correct superpositions is not trivial and my algorithm for that is not perfect for sure.
However, I'm still a bit surprised that it fails for your model because you don't have any other combinations than a THDM-III where it works...

Cheers,
Florian

Hi,

I attach a first update of the routines. Unfortuantely, it still throws erros concerning lambda5 and lambda7. I guess the reason is that SARAH doesn't derive the symmetries of the quartics, yet. I hope I can continue looking at that issue tomorrow.

Cheers,
Florian