This repository contains a C++ library for finding affine decision trees for vectorial Boolean functions and is published alongside the 2021 Asiacrypt paper "Generic Framework for Key-Guessing Improvements".
We tested the library with the following dependencies on a Linux machine:
| Dependency | Version tested |
|---|---|
| [gcc] (https://gcc.gnu.org) | 11.1.0 |
| [cmake] (https://cmake.org) | 3.21.2 |
| [NTL] (https://libntl.org) | 11.5.2 |
| [GMP] (https://gmplib.org) | 6.2.1 |
| [openmp*] (https://openmp.llvm.org/) | 12.0.1 |
| [doxygen**] (https://www.doxygen.nl) | 1.9.2 |
where the dependencies marked with a * are only necessary for compiling all of the examples, those with ** are only necessary for building the documentation. We make use of gcc's C++20-support (mainly for std::popcount).
The library needs to be compiled (i.e. it is not header only). It suffices to run cmake on the top-level CMakeLists.txt and then use the resulting makefile for building the library.
cd <build directory>
cmake <project's root director>
make all
You can install the library system-wide by running make install with the appropriate privileges. The exported
target for CMake is ConditionsLib::ConditionsLib.
For an example how to use the library with CMake refer to examples-directory.
A pdf file of the code documentation can be found in the doc/-folder.
To build the documentation, run
cd <documentation directory>
cmake <project's root directory>
make doxygen
The equivalence classes of balanced functions up to dimension 6 and 4-bit Sboxes in the FunctionData directory were given to us by Shahram Rasoolzadeh.