I haven't actively maintained GENEUS for a while, and unfortunately it is currently in a broken state. It doesn't compile cleanly under the last versions of GFortran or Intel Fortran, and often crashes due to unidentified segfaults when compiled using old compiler versions that used to work. I would therefore not recommend using it unless you're comfortable debugging Fortran yourself.
One alternative is to look at this older Usadel solver; it's much slower and has fewer features, but last I checked the code still worked. Please also check out my newer Bogoliubov-de Gennes package for Python if you want to model superconductivity in the clean limit.
You need a Linux/Unix system with GFortran v8.x and CMake v3.x. Note that the code is susceptible to segfaults when compiled with newer GFortran versions, and might not compile at all using older versions or other compilers.
The code can be compiled as follows:
./configure
cd build
make
The executables will then end up in the project subfolder bin. See the documentation for more information on how to use these binaries for physics simulations.
Before running any simulations, it's recommended that you add
ulimit -s unlimited
to your shell config (e.g. ~/.bashrc). If not, limitations in stack size may result in crashes if you attempt to multiple simulations in parallel.
In contrast to GENEUS v1, the current version only supports a single cmake build type, a single platform (Ubuntu x86-64), and a single compiler (GFortran). Narrowing down the debugging to one compilation target simplifies maintenance. The binaries are now also portable (fully static x86-64 binaries), meaning that
they can be transferred between computers without requiring recompilation.
One drawback of the above is performance. By dropping support for Intel Fortran, and switching to less aggressive compilation flags, GENEUS v2 runs up to an order of magnitude slower than v1 did. If this becomes a bottleneck, you are of course free to experiment with other compilation options; please don't be shy to reach out with potential bugfixes and optimizations in that case.
I've attempted to make the code conform more to the conventions used in modern Fortran code, e.g. standard file extensions and line lengths. These updates were performed with the help of fprettify. Hopefully, this should make the code easier to extend by others if needed.