FlexibleSUSY provides Mathematica and C++ code to create spectrum generators for supersymmetric and non-supersymmetric models. It is designed for generating fast and modular C++ code, allowing for easy modification, extension and reuse.
Homepage: https://flexiblesusy.hepforge.org
Mailing list: [email protected]
Source code repository: https://github.com/FlexibleSUSY
Bug reports: https://github.com/FlexibleSUSY/FlexibleSUSY/issues
References: [1406.2319], [1609.00371], [1710.03760]
If you use FlexibleSUSY in your work please cite [1406.2319] and [1710.03760].
If you use the FlexibleEFTHiggs approach in your work please cite [1609.00371] and [1710.03760].
If you use FlexibleSUSY+Himalaya or Himalaya in your work, please cite [1005.5709], [1708.05720], [1807.03509] and [1910.03595].
FlexibleSUSY depends on SARAH and contains components from SOFTSUSY. Therefore, please also cite the following publications along with FlexibleSUSY:
- SARAH [0909.2863], [1002.0840], [1207.0906], [1309.7223]
- SOFTSUSY [hep-ph:0104145], [1311.7659]
The list of references in BibTeX format can be found in doc/references.bib.
Table of Contents
- 1 Quick start
- 2 Building FlexibleSUSY
- 3 Using FlexibleSUSY
- 4 Advanced FlexibleSUSY build options
- 4.1 Generating source code files only (no compilation)
- 4.2 Compile only (don't generate source code)
- 4.3 Exporting the generated source code
- 4.4 Dynamic libraries
- 4.5 Statically linked executables
- 4.6 LoopTools support
- 4.7 TSIL support
- 4.8 Creating an addon
- 4.9 Creating the source code documentation
- 4.10 Cleaning
- 5 Package content
- 6 Further reading
- 7 References
Install required libraries and packages (if not already done):
pip install conan conan remote add conan-hep https://api.bintray.com/conan/expander/conan-hep conan install . --build=missing sudo apt-get install libgsl-dev ./install-sarah
Build a spectrum generator (here: HSSUSY [1710.03760] [1804.09410]):
./createmodel --name=HSSUSY ./configure --with-models=HSSUSY make -j4
Run the spectrum generator:
./models/HSSUSY/run_HSSUSY.x --slha-input-file=model_files/HSSUSY/LesHouches.in.HSSUSY
- C++ compiler (g++ >= 5.0.0 or clang++ >= 3.8.1 or icpc >= 17.0.0)
- Fortran compiler (gfortran, ifort)
- Mathematica (version 7.0 or higher)
- SARAH (version 4.11.0 or higher)
- Boost (version 1.37.0 or higher)
- Eigen 3 (version 3.1 or higher)
- GNU scientific library
Optional:
- FeynArts (version 3.9 or higher)
- FormCalc (version 9.5 or higher)
- LoopTools (version 2.8 or higher)
- Himalaya
- TSIL
The required libraries Boost, Eigen 3, LoopTools, Himalaya and TSIL can be installed using the Conan package manager. If not already installed, Conan can be installed with pip:
pip install conan
If Conan is installed, add conan-hep to the list of remote package repositories:
conan remote add conan-hep https://api.bintray.com/conan/expander/conan-hep
To install the libraries required by FlexibleSUSY run:
conan install . --build=missing
The GNU scientific library can currently not be installed via Conan. One may use the package manager of the operating system to install it. On Debian/Ubuntu one may run for example:
sudo apt-get install libgsl-dev
If the required libraries are installed via Conan or the operating
system's package manager, they will be found automatically by
FlexibleSUSY's configure
script, see below.
FlexibleSUSY requires SARAH to be installed and to be loadable with
the Needs["SARAH`"]
command from inside Mathematica. We recommend
the following setup:
SARAH_VERSION=4.14.3 cd ~/.Mathematica/Applications/ wget https://sarah.hepforge.org/downloads/SARAH-${SARAH_VERSION}.tar.gz tar -xf SARAH-${SARAH_VERSION}.tar.gz ln -s ${PWD}/SARAH-${SARAH_VERSION}/ SARAH cd ~/.Mathematica/Kernel/ echo "AppendTo[\$Path, \"${HOME}/.Mathematica/Applications/SARAH/\"];" >> init.m
All the above steps can be executed at once with the install-sarah
script:
./install-sarah
See ./install-sarah --help
for more options.
If you want FlexibleSUSY to use FeynArts or FormCalc you will need
to install these packages first. Also — as with SARAH — they need to
be loadable with the Needs[]
command from inside Mathematica. We
recommend using the installation script FeynInstall
provided on
the FeynArts web page. e.g.:
cd ~/.local wget http://www.feynarts.de/FeynInstall chmod 755 FeynInstall ./FeynInstall
which will install the latest versions of FeynArts, FormCalc and
LoopTools in the ~/.local/
directory as well as configure
Mathematica to find these packages. Note that running the
FeynInstall
script might require user intervention.
Before you setup a FlexibleSUSY model, you have to provide a SARAH model file. To make it available in FlexibleSUSY, you can put it either into FlexibleSUSY's SARAH model directory
FlexibleSUSY/sarah/<model>/
or directly into SARAH's own model directlySARAH/Models/<model>/
. Here<model>
is the name of your model (e.g. MSSM, NMSSM, etc.). Note, that there are already plenty of pre-installed model files in FlexibleSUSY's and SARAH's model directories that can be used.Create a new or re-initialize an existing FlexibleSUSY model:
./createmodel --name=<model>
See
./createmodel --help
for more details. Afterwards there will be- a model directory
models/<model>/
- a makefile module
models/<model>/module.mk
- a Mathematica start script
models/<model>/start.m
- and a FlexibleSUSY model file
models/<model>/FlexibleSUSY.m
To modify the model details (input parameters, boundary conditions, etc.), edit the FlexibleSUSY model file
models/<model>/FlexibleSUSY.m
. For more details see the documentation of the FlexibleSUSY model file and FlexibleEFTHiggs.- a model directory
Create the Makefile and register your model(s):
./configure --with-models=<model>
Multiple models can be specified, separated by a comma. See
./configure --help
for more options.Compile FlexibleSUSY with your model:
make
Use
make -j<N>
to use<N>
CPU cores. Whenmake
is executed, Mathematica is called, which generates the C++ code for the specified models. All C++ source files are written to the directorymodels/<model>/
. Whenmake
has finished, the following spectrum generator(s) are available for each specified model:models/<model>/run_<model>.x
: command line spectrum generatormodels/<model>/run_<model>.m
: Mathematica interface
Example:
./createmodel --name=HSSUSY ./configure --with-models=HSSUSY make -j4 ./models/HSSUSY/run_HSSUSY.x --slha-input-file=model_files/HSSUSY/LesHouches.in.HSSUSY
FlexibleSUSY ships with many pre-generated models. The following table includes an (incomplete) list of models with a detailed documentation.
Model | Description |
---|---|
HSSUSY | high-scale MSSM (pure EFT) |
MSSMEFTHiggs | high-scale MSSM (FlexibleEFTHiggs) |
NUHMSSMNoFVHimalaya | fixed-order MSSM |
For each model FlexibleSUSY creates an executable
models/<model>/run_<model>.x
that can be run from the command
line. The executable accepts the input in the SLHA format, for
example in form of a file:
./models/MSSM/run_MSSM.x \ --slha-input-file=models/MSSM/LesHouches.in.MSSM \ --slha-output-file=LesHouches.out.MSSM
or as a stream:
cat models/MSSM/LesHouches.in.MSSM \ | ./models/MSSM/run_MSSM.x --slha-input-file=- --slha-output-file=LesHouches.out.MSSM
For a documentation of FlexibleSUSY-specific switches in the SLHA input see the section on SLHA input parameters.
By default the executable writes the output in SLHA format to stdout. The output can also be appended to an SQLite database:
./models/MSSM/run_MSSM.x \ --slha-input-file=models/MSSM/LesHouches.in.MSSM \ --slha-output-file=LesHouches.out.MSSM \ --database-output-file=points.db
See models/<model>/run_<model>.x --help
for further options.
The pole mass spectrum and the RG flow can be written to text files for easy plotting. In the MSSM for example these text files can be generated via:
./models/MSSM/run_MSSM.x \ --slha-input-file=model_files/MSSM/LesHouches.in.MSSM \ --rgflow-output-file=MSSM_rgflow.dat \ --spectrum-output-file=MSSM_spectrum.dat
The generated files MSSM_rgflow.dat
and MSSM_spectrum.dat
can
be plotted for example with the gnuplot scripts in the model
directory:
gnuplot -persist -e "filename='MSSM_spectrum.dat'" \ models/MSSM/MSSM_plot_spectrum.gnuplot gnuplot -persist -e "filename='MSSM_rgflow.dat'" \ models/MSSM/MSSM_plot_rgflow.gnuplot
The gnuplot scripts are just for illustration and currently plot all running parameters, regardless of their mass dimension, so the resulting plot is not particularly informative. However, one may easily adapt the scripts to plot any chosen subset of the parameters.
FlexibleSUSY can be called from within Mathematica using Wolfram's
LibraryLink. By default, FlexibleSUSY creates a LibraryLink library
for each spectrum genreator. The generated library can be found in
models/<model>/<model>_librarylink.so
, where <model>
is the
model name.
Example:
Get["models/CMSSM/CMSSM_librarylink.m"]; (* Create a handle to a model given the input parameters. See Options[FSCMSSMOpenHandle] for all default options. *) handle = FSCMSSMOpenHandle[ fsSettings -> { precisionGoal -> 1.*^-4 }, fsSMParameters -> { Mt -> 173.3 }, fsModelParameters -> { m0 -> 125, m12 -> 500, TanBeta -> 10, SignMu -> 1, Azero -> 0 } ]; (* calculate pole mass spectrum *) FSCMSSMCalculateSpectrum[handle] (* calculate observables *) FSCMSSMCalculateObservables[handle] (* close the model handle *) FSCMSSMCloseHandle[handle];
For each model, FlexibleSUSY creates an example Mathematica script
which illustrates the use of the Mathematica interface. The generated
example can be found in models/<model>/run_<model>.m
which can be
run for example as:
math -run "<< \"models/<model>/run_<model>.m\""
Before running it, the model parameters in the script should be set to reasonable values. More advanced examples can be found in the FlexibleSUSY documentation.
Note: In order to compile the library, Mathematica must be installed.
To disable the LibraryLink interface, configure with
--disable-librarylink
.
Further details and examples can be found in the LibraryLink documentation.
FlexibleSUSY contains two shell scripts aiming to help the user performing parameter scans based on SLHA files.
The script utils/scan-slha.sh
performs a scan over an input
parameter.
Examples:
To perform a scan over \tan\beta(M_Z) in the CMSSM (given in
the SLHA input file in the MINPAR[3]
field) and print out the the
values of \tan\beta(M_Z), M_h (MASS[25]
) and
y_t(M_{\text{SUSY}}) (YU[2,2]
) run:
utils/scan-slha.sh \ --spectrum-generator=models/CMSSM/run_CMSSM.x \ --slha-input-file=model_files/CMSSM/LesHouches.in.CMSSM \ --scan-range=MINPAR[3]=1~30:10 \ --output=MINPAR[3],MASS[25],YU[2:2]
Alternatively, the SLHA input can be piped into the script as
cat model_files/CMSSM/LesHouches.in.CMSSM \ | utils/scan-slha.sh \ --spectrum-generator=models/CMSSM/run_CMSSM.x \ --scan-range=MINPAR[3]=1~30:10 \ --output=MINPAR[3],MASS[25],YU[2:2]
The spectrum generator executable is specified using the
--spectrum-generator=
option. The parameter to be scanned over as
well as the scan range and the number of steps must be specified using
the --scan-range=
option. The syntax is:
--scan-range=<block>[<field>]=<start>~<stop>:<number_of_steps>
Here <block>
is the SLHA block in which the input parameter is to
be found and <field>
is the block entry corresponding to the
parameter. <start>
and <stop>
define the scan range and
<number_of_steps>
define the number of steps. By default the step
size is linear. Alternatively, a logarithmic step size can be chosen
by passing --step-size=log
to the script. See also
utils/scan-slha.sh --help
. The parameters to print to the output
stream must be defined using the --output=
option. The syntax
is:
--output=<block>[<fields>]
where <block>
is the SLHA block in which the output parameter is to
be read from and <field>
is the block entry corresponding to the
parameter. To read a matrix element from a block, use a colon :
to
specify the matrix element indices. Multiple output parameters can be
specified by a comma.
As an alternative, all parameters calculated during a scan can be
written to a SQLite database using the scan-database.sh
script.
Examples:
utils/scan-database.sh \ --spectrum-generator=models/CMSSM/run_CMSSM.x \ --slha-input-file=model_files/CMSSM/LesHouches.in.CMSSM \ --scan-range=MINPAR[3]=1~30:10 \ --database-output-file=scan.db
or:
cat model_files/CMSSM/LesHouches.in.CMSSM \ | ./utils/scan-database.sh \ --spectrum-generator=models/CMSSM/run_CMSSM.x \ --scan-range=MINPAR[3]=1~30:10 \ --database-output-file=scan.db
The name of the database file must be set using the
--database-output-file=
option.
The script utils/convert_SPheno_to_FlexibleSUSY.m
can help to
convert a SPheno model file (SPheno.m
) to a FlexibleSUSY model
file (FlexibleSUSY.m.in
). The conversion is not perfect, because
it is usually not unique. Therefore one should check the generated
FlexibleSUSY.m.in
file.
Example:
cat << EOF | math -noprompt > FlexibleSUSY.m.in sphenoFile = "~/.Mathematica/Applications/SARAH/Models/MSSM/SPheno.m"; Get["utils/convert_SPheno_to_FlexibleSUSY.m"]; EOF
If you want to only create the C++ source files for your model, but do
not want to compile the code, you can use the --disable-compile
configure option:
./configure --with-models=MSSM --disable-compile make
Here, configure will not check for installed compilers or libraries.
It will only search for Mathematica and SARAH. The execution of
make
will stop as soon as all C++ source code files are generated.
See below for how to export the generated source code.
If you want to only compile already created the C++ source files for
your model, you can use the --disable-meta
configure option:
./configure --with-models=MSSM --disable-meta make
Here, configure will only check for installed compilers or libraries. It will not check for Mathematica and SARAH.
This option is useful if you want to generate the source code on one computer and then transfer the generated code to another computer to compile it. This option can also be used with the pre-generated FlexibleSUSY models, which are provided at the FlexibleSUSY home page.
Warning: Please make sure all C++ source files of your model are
available in the model directory models/<model>/
. Otherwise the
compilation will fail.
The complete FlexibleSUSY source code, including the generated C++
code for the specified model(s) (but without the Mathematica meta
code), can be exported to a new directory. The exported source code
is a complete standalone package, with it's own build system. To
export the code, one has to set the target directory during
configuration via the --with-install-dir=
option. For example:
./configure --with-models=<models> --with-install-dir=/path/to/export/directory
Afterwards
make install-src
must be executed, which will copy the generated C++ source code for
all <models>
to /path/to/export/directory
, together with the
non-model specific source code from config/
, doc/
, slhaea/
and src/
. Afterwards, the standalone package can be build like
this:
cd /path/to/export/directory ./configure make
It is also possible to create a "model package", which includes only the generated source code for a given model, but does not contain the whole FlexibleSUSY build system. This is useful when the source code for a model should be generated on one computer and later transferred to another one to be compiled. To create such a "model package" run
make pack-<model>-src
where <model>
is the name of the model whose generated source code
shall be packed. After make
has finished, the package file
<model>.tar.gz
can be found in the working directory.
If you want to create dynamic model libraries (instead of static
libraries, which is the default) you need to pass the
--enable-shared-libs option to the configure script. The file name
extension for the shared libraries as well as the command to build
them can be overwritten using the --with-shared-lib-ext=
--with-shared-lib-cmd=
. parameters. For example, when Intel
compilers should be used, replace gcc by icc or icpc:
./configure --with-models=CMSSM,NMSSM \ --enable-shared-libs \ --with-shared-lib-ext=".so" \ --with-shared-lib-cmd="gcc -shared -o"
Important remark:
The libraries are linked to the executables with absolute paths. This means that, if you for example move the FlexibleSUSY directory to another location, the executables will no longer find the libraries. To make the executables find the libraries again, you have to relink them via
make clean-executables make allexec
External libraries can be linked statically to the spectrum generator
executables by passing --enable-static
to configure. This is
useful when the executable should be transferred to another computer,
where some libraries are not available.
Example:
./configure --with-models=CMSSM --enable-static
If --enable-static
is used, the following linker flags and
additional libraries will be used:
LDFLAGS = -static LDLIBS = -ldl
These linker-specific flags and additional libraries can be
overwritten using --with-static-ldflags=
and
--with-static-ldlibs=
Example:
./configure --with-models=CMSSM \ --enable-static \ --with-static-ldflags="-static" \ --with-static-ldlibs="-lquadmath -ldl"
In case of dynamic linking (--disable-static
, which is the default),
the options --with-shared-ldflags=
and --with-shared-ldlibs=
must
be used to set LDFLAGS
and LDLIBS
.
It is possible to use LoopTools for calculating the loop functions, instead of using SOFTSUSY's loop functions. To enable LoopTools configure FlexibleSUSY via
./configure --enable-looptools
To use the LoopTools library and header files from a specific directory configure via
LOOPTOOL_DIR=/path/to/looptools/build ./configure --enable-looptools \ --with-looptools-incdir=$LOOPTOOLS_DIR \ --with-looptools-libdir=$LOOPTOOLS_DIR
Note: LoopTools 2.8 or higher is required.
Some models of FlexibleSUSY require TSIL, for example HSSUSY. When
such models are activated (via ./configure --with-models=<model>
),
FlexibleSUSY requires TSIL to be available. If TSIL is installed in a
system directory or installed via Conan, FlexibleSUSY will find the
TSIL automatically. To use TSIL from a a non-standard directory,
configure FlexibleSUSY like this:
$TSIL_DIR=/path/to/tsil ./configure --enable-tsil \ --with-tsil-incdir=$TSIL_DIR \ --with-tsil-libdir=$TSIL_DIR
Note also that TSIL must be compiled with -fPIC
, which can be
achieved by setting in the TSIL Makefile
:
TSIL_OPT = -O3 -funroll-loops -fPIC
A FlexibleSUSY addon is a program or library, which uses parts of the FlexibleSUSY libraries or the generated models or is integrated into FlexibleSUSY. An example is GM2Calc, which is included in FlexibleSUSY in form of an addon. An addon can be created via
./createaddon --name=<addon>
where <addon>
is the name of the addon. The createaddon script
creates the directory addons/<addon>/
and the corresponding makefile
module addons/<addon>/module.mk
. If an addon has been created with
the above script, the user may edit the makefile module
(addons/<addon>/module.mk
) to add source files in to the three
variables
LIB@ADDON@_SRC # list of source files to be included in library EXE@ADDON@_SRC # list of source files with a main() LIB@ADDON@_HDR # list of header files
Example:
LIB@ADDON@_SRC := $(DIR)/file1.cpp EXE@ADDON@_SRC := $(DIR)/run.cpp LIB@ADDON@_HDR := $(DIR)/file1.hpp
To configure and compile the addon run
./configure --with-addons=<addon> make
make compiles all source files and creates the addon library
addons/<addon>/lib<addon>.a
(including the object file file1.o
in
the above example) and an executable (addons/<addon>/run.x
in the
above example).
FlexibleSUSY's source code documentation (including the generated source code files) can be generated with Doxygen in HTML or man format. To generate the HTML documentation please run:
make doc-html
The generated HTML index file can then be found in
doc/html/index.html
and can be viewed with any HTML browser, e.g.
firefox doc/html/index.html
To generate the man documentation please run:
make doc-man
The generated man pages can then be found in doc/man/man3/
and can
be viewed as
man doc/man/man3/model_file_options.3
There are several make targets to remove generated files, compiled object files, libraries or executables:
make clean # deletes all .d .o .a .x files make distclean # does `clean` and `clean-generated` # and deletes in addition: # Makefile flexiblesusy-config config.* # config/list_sarah_model_files.sh make clean-dep # deletes all .d files make clean-executables # deletes all .x files make clean-generated # deletes generated files make clean-lib # deletes all libraries make clean-obj # deletes all .o files
For each model <model>
or addon there are specific clean targets
to remove model-specific files:
make clean-<model> # deletes .d .o .a .x files make distclean-<model> # same as `make clean-<model> clean-<model>-src` make clean-<model>-dep # deletes .d files make clean-<model>-lib # deletes model library make clean-<model>-obj # deletes .o files make clean-<model>-src # deletes generated files
In the following all sub-directories within the FlexibleSUSY package are listed:
addons/
contains addons for FlexibleSUSY, such as GM2Calcconfig/
contains helper scripts and makefile modules for the build systemdoc/
contains the FlexibleSUSY documentationexamples/
contains examples how to build you own spectrum generator based on FlexibleSUSYfflite/
contains an alternative implementation of the Passarino-Veltman loop functions, based on FFmeta/
contains the Mathematica meta code which generates the spectrum generators. See the meta code documentation for more details.model_files/
contains default model files for some frequently used models (SM, SplitMSSM, MSSM, NMSSM, SMSSM, UMSSM, etc.)model_specific/
contains model-specific higher order corrections for the MSSM, NMSSM, SM and SplitMSSM from the literaturemodels/
This is the output directory where the generated C++ code for the spectrum generators will be stored.Output/
contains SARAHs model-specific output filessarah/
contains SARAH model files shipped with FlexibleSUSYslhaea/
contains the slhaea SLHA reader librarysrc/
contains model-independent FlexibleSUSY C++ source codetemplates/
contains C++ template files for the spectrum generatorstest/
contains the FlexibleSUSY test suiteutils/
contains some utility scripts to perform scans or extract data from SLHA files
- FlexibleSUSY model file
- FlexibleEFTHiggs
- LibraryLink documentation
- meta code documentation
- SLHA input parameters
[hep-ph:0104145] | CPC 143 (2002) 305-331 [arxiv:hep-ph/0104145] |
[0909.2863] | CPC 181 (2010) 1077-1086 [arxiv:0909.2863] |
[1002.0840] | CPC 182 (2011) 808-833 [arxiv:1002.0840] |
[1005.5709] | JHEP 1008 (2010) 104 [arxiv:1005.5709] |
[1207.0906] | CPC 184 (2013) 1792-1809 [arxiv:1207.0906] |
[1309.7223] | CPC 185 (2014) 1773-1790 [arxiv:1309.7223] |
[1311.7659] | CPC 185 (2014) 2322 [arxiv:1311.7659] |
[1406.2319] | (1, 2) CPC 190 (2015) 139-172 [arxiv:1406.2319] |
[1609.00371] | (1, 2) JHEP 1701 (2017) 079 [arxiv:1609.00371] |
[1708.05720] | Eur.Phys.J. C77 (2017) no.12, 814 [arxiv:1708.05720] |
[1710.03760] | (1, 2, 3, 4) CPC 230 (2018) 145-217 [arXiv:1710.03760] |
[1804.09410] | Eur.Phys.J. C78 (2018) no.7, 573 [arxiv:1804.09410] |
[1807.03509] | Eur.Phys.J. C78 (2018) no.10, 874 [arxiv:1807.03509] |
[1910.03595] | Eur.Phys.J. [arxiv:1910.03595] |