If you think you may have found a bug in FlexibleSUSY, which is not listed below, please file a bug report.
If FlexibleSUSY exits with an exception indicating that threads cannot
be opened (which happens for example with g++ 4.8.x), one can try
adding -Wl,--no-as-needed to the CXXFLAGS, i.e. for example:
./configure --with-cxxflags="-Wl,--no-as-needed -std=c++11 -O2"
If this does not solve the problem, one can disable multi-threading via:
./configure --disable-threads
FlexibleSUSY cannot yet deal with tensor couplings (matrix-like couplings with rank 3, i.e. with 3 or more indices) in the Lagrangian/Superpotential. In order to make such models work in FlexibleSUSY, please split the tensor coupling terms in the Lagrangian/Superpotential into a sum of terms with a matrix-like coupling structure of rank 2 or less.
FlexibleSUSY cannot yet include all effects from fermions, which mix with the Standard Model fermions, in the extraction of the running Yukawa couplings. In such models, the Yukawa couplings have to be fixed by hand in one boundary condition. FlexibleSUSY provides the symbols
upQuarksDRbar= (mu, ms, mt)downQuarksDRbar= (md, mc,mb)downLeptonsDRbar= (me, mµ, mτ)
to access the running Standard Model fermion masses in such models. See the munuSSM model file for an example, where the charginos mix with the Standard Model fermions. In this model the down-type lepton Yukawa matrix is fixed in an approximated form at the low-energy scale given the running SM charged lepton masses.
The Tensor module of the Eigen library before version 3.3.5 may
fail to compile with clang++ due to a bug in Eigen. The bug has
been fixed in Eigen 3.3.5.
When models with complex parameters are built using versions of Eigen 3.3.0, it may be the case that compilation of the generated code fails with an error message indicating that different numeric types have been mixed without using an appropriate cast. This arises because of limitations in mixing different scalar types in earlier versions of Eigen. In some cases, this problem can be worked around by reformulating the boundary conditions in the model file. For example, in a model such as the MSSMCPV, a boundary condition of the form:
{T[Ye], (ReAeInput + I ImAeInput) Ye},
{T[Yd], (ReAdInput + I ImAdInput) Yd},
{T[Yu], (ReAuInput + I ImAuInput) Yu}
may trigger this compilation error when compiled against versions of Eigen below 3.3.0, due to the terms in brackets involving addition of a real and complex matrix. In cases such as this, the problem can be avoided by rewriting the boundary condition to avoid an operation involving mixed matrix types. For example, in the equivalent form:
{T[Ye], ReAeInput Ye + I ImAeInput Ye},
{T[Yd], ReAdInput Yd + I ImAdInput Yd},
{T[Yu], ReAuInput Yu + I ImAuInput Yu}
all additions involve only complex matrices.
If it is not possible to solve this compilation error by reformulating the boundary conditions of the model, it is recommended that the installed Eigen version be upgraded to at least 3.3.0, in which this problem no longer occurs due to better support for operations involving mixed numerical types.
It can happen that Mathematica crashes on exit when multiple Mathematica interfaces of FlexibleSUSY (LibraryLinks) were loaded. The crash occurs only when Mathematica is closed, not during run-time. Depending on the Mathematica version and the used system, the crash could be caused by one of the following problems:
*** Error in ``[...]/WolframKernel': corrupted size vs. prev_size *** Error in ``[...]/WolframKernel': corrupted double-linked list *** Error in ``[...]/WolframKernel': free(): invalid pointer *** Error in ``[...]/MathKernel': double free or corruption (!prev)
In some cases the crash could be avoided by disabling multi-threading
(see ./configure --help) or by using a different Mathematica
version.
When such a crash occurs during run-time, please file a bug report.