The most recent version of LLAMA can be found at GitHub.
git clone https://github.com/alpaka-group/llama cd llama
All examples use CMake and the library itself provides a
llama-config.cmake to be found by CMake.
Although LLAMA is a header-only library, it provides installation capabilities via CMake.
At its core, using the LLAMA library requires:
cmake 3.18.3 or higher
Boost 1.74.0 or higher
libfmt 6.2.1 or higher (optional) for support to dump mappings as SVG/HTML
Building the unit tests additionally requires:
Catch2 3.0.1 or higher
To build all examples of LLAMA, the following additional libraries are needed:
Build tests and examples¶
As LLAMA is using CMake the tests and examples can be easily built with:
mkdir build cd build cmake .. -DBUILD_TESTING=ON -DLLAMA_BUILD_EXAMPLES=ON ccmake .. // optionally change configuration after first run of cmake cmake --build .
This will search for all dependencies and create a build system for your platform. If necessary dependencies are not found, the corresponding examples will be disabled. After the initial call to cmake, ccmake can be used to add search paths for missing libraries and to deactivate building tests and examples.
To install LLAMA on your system, you can run (with privileges):
cmake --install .