Supplementary Data: Global fits of GUT-scale SUSY models with GAMBIT (arXiv:1705.07935)

Supplementary Data

Global fits of GUT-scale SUSY models with GAMBIT
arXiv:1705.07935

The files in this record contain data for the CMSSM, NUHM1 and NUHM2 models considered in the GAMBIT "Round 1" GUT-scale SUSY paper.

For each model, there are

• A number of YAML files, each corresponding to a different set of sampling parameters and/or priors
• A set of YAML files used for postprocessing: CMSSM_intermediate.yaml, CMSSM.yaml, NUHM1.yaml and NUHM2.yaml
• A final hdf5 file, containing the combined results of all sampling runs
• An example pip file, for producing plots from the hdf5 file using pippi
• SLHA1 and SLHA2 files for the best-fit point in each subregion of the fit. These can be found inside the tarball best_fits_SLHA.tar.gz.

The record also contains

• StandardModel_SLHA2_scan.yaml and StandardModel_SLHA2_postprocessing.yaml, two universal YAML fragments included from other yaml files
• gambit_preamble.py, a collection of python functions used for in-line data processing in the pip files

The different YAML files corresponding to different samplers and/or priors follow the naming scheme [model]_[scanner]_[prior]_[slice]_[special].yaml, where

• model = CMSSM, NUHM1, NUHM2
• scanner = Diver, MN
• prior = log, flat
• slice = pmu, nmu (positive or negative mu)
• special = sqcoann, slcoann, [blank] (squark co-annihilation, slepton co-annihilation, or bulk)

A few caveats to keep in mind:

1. For each model, the final hdf5 results file included here was generated in the following way:

   • carry out initial runs using YAML files following the naming scheme above
   • combine the resulting hdf5 output files into a single file, using gambit/Printers/scripts/combine_hdf5.py
   • postprocess the samples to remove all points more than 5 sigma from the current best fit, using [model]_strip.yaml
   • postprocess the samples to include a new likelihood term for LHC Run II searches, and to recompute the FlavBit likelihoods (these were buggy in a pre-release version of GAMBIT). For the CMSSM, this happened in two steps, due to persistent flavour bugs, using CMSSM_intermediate.yaml and CMSSM.yaml. For the NUHM1 and NUHM2, this was done in a single step each, using NUHM1.yaml and NUHM2.yaml.

2. It is not necessary to repeat the steps listed in point 1 when running new scans; the LHC Run II likelihoods can be included in the original YAML file, so that no postprocessing step is required.

3. The YAML files that we give here are updated compared to the ones that we used when generating the hdf5 file, in order to match the set of available options in the release version of GAMBIT 1.0.0. The included physics and numerics are however identical.

4. The YAML files are designed to work with the tagged release of GAMBIT 1.0.0, and the pip files are tested with pippi 2.0, commit 2ab061a8. They may or may not work with later versions of either software (but you can of course always obtain the version that they do work with via the git history).

5. The pip file for each model is an example only. Users wishing to reproduce the more advanced plots in any of the GAMBIT papers should contact us for tips or scripts, or experiment for themselves. Many of these scripts are in multiple parts and require undocumented manual interventions and steps in order to implement various plot-specific customisations, so please don't expect the same level of polish as for files provided here or in the GAMBIT repo.