SSP and micromegas

[Amin]

Hello experts,

I am using MCMC in SSP for the MSSM that comes with micromegas (not the SARAH generated one). It seems to me that it doesn't work since I am getting "-1 Wrong LSP" in the "MCMCInfo" block. When I run the file "SPheno.spc" manually with micromegas "main" function it works (after commenting out items 2 and 6 in the MODSEL block) but it looks like calling it with SSP fails. Is that true? SSP can only work with SARAH generated models?
I get this error in mathematica as well:

Power::infy: Infinite expression 1/0 encountered.
Infinity::indet: Indeterminate expression 0 ComplexInfinity encountered.
Part::partd: Part specification Null[[2]] is longer than depth of object.
Part::partd: Part specification Null[[2]] is longer than depth of object.
Found valid starting point. Likelyhood: Null[[2]]
Power::infy: Infinite expression 1/0 encountered.
Infinity::indet: Indeterminate expression 0 ComplexInfinity encountered.
Part::partd: Part specification Null[[2]] is longer than depth of object.
General::stop: Further output of Part::partd will be suppressed during this calculation.

Of course I did check whether my assignment of the LSP is wrong but it isn't. 1000022 is the LSP in my case.
One more thing: is it possible to parallelize MCMC in SSP?

Thank you,
Amin

[FStaub]

Hi,

SSP doesn't need necessarily a SARAH model file. However, the main file for micromegas must be adjusted because SSP expects a file "omg.out" where the relic density and the annihilation channels are given. See for instance the CalcOmega.cpp written by SARAH.

Well, you can just start different MCMC using different mathematica kernels. That's the best parallelisaiton which is possible so far. However, you need to be careful: MO works with absolute paths, i.e. each chain needs it its own MO installation to prevent conflicts.

Cheers,
Florian

[Amin]

Hi Florian,

Thanks for the info.
In case I wanted to use MSSM from SARAH rather than the one that comes with SPheno and micromegas, is it fair to ask which one is better or which has more accuracy? I found some differences in the output between the two.

Best,
Amin

[FStaub]

Hi,

there shouldn't be any significant difference. Both combinations calculate the masses at the same loop-level and the relic density at tree-level. Any difference should just be due to some different input choices for SM parameters or so.

Cheers,
Florian

[Amin]

Yes there are differences in the SM input. I suppose this accounts for the difference I am noticing between the two.

Thank you!!
Amin