New recommended sampling parameters for accurate mapping of the profile likelihood with MultiNest. See 1101.3296 for details.
New version v1.5.1 now available (April 2011).
here to see the new
features. Some changes to the MultiNest sampler to make it more robust on different
architectures. There might be some sampling issues with the previous version (v1.5) on some architectures, so we recommend upgrading to v1.5.1 and checking the accuracy of the samples obtained with v1.5.
Online plotting routine now available
SuperBayeS is a package for fast and efficient
supersymmetric theories. It uses
Bayesian techniques to explore multidimensional SUSY parameter
spaces and to compare SUSY predictions with observable
quantities, including sparticle masses, collider observables,
dark matter abundance, direct detection cross sections,
indirect detection quantities etc. Scanning can be performed
using Markov Chain Monte Carlo (MCMC) technology or even more
efficiently by employing a new scanning technique called,
which implements the
nested sampling algorithm. Using MultiNest, a full
8-dimensional scan of the CMSSM takes about 12 hours on
10 2.4GHz CPU's. There is also an option for old-style
More info about the package can be found throughout this
If you discover any bug or if you have any questions,
please login on the SuperBayeS discussion
The package combines
MicrOMEGAs. Some of the routines and the plotting tools are based
SuperBayeS now comes with
SuperEGO, a MATLAB graphical user interface tool for interactive plotting of the results. SuperEGO has been developed by Rachid Lemrani and is based on CosmoloGUI by Sarah Bridle.
Users are kindly requested to add the following statement
(or something to the same effect) to their acknowledgements
in any publication using SuperBayeS:
The author(s) acknowledge the use of the SuperBayeS package,
which includes the independently developed codes SoftSusy,
DarkSUSY, MicrOMEGAs, FeynHiggs, Bdecay and MultiNest
and which employs some of the CosmoMC package routines.
If you use the package, please acknowledge the following
Also, you might want to read the original papers
where the method is described in some detail.