This is the documentation for CudaTools, a header-only library and framework for the development of CPU-CUDA compatible applications. Using CudaTools enables the creation of a single unified code that has both CPU and CUDA compilation targets with minimal need to introduce #ifdef statements when code is essentially identical between the targets.

For information on the library itself and its usage, view documentation. The small code snippets and samples seen in the documentation are in the folder samples.

Dependencies

Eigen

In the future, we will make this dependency optional, but still provide support for it. As of now, it is necessary.

Building the Documentation

The documentation is built with Doxygen and Sphinx. So, first make sure you have Doxygen installed on your system, and make sure it is added to your system path. Then, you will have to create a Python virtual environment in the repository folder

$ python3 -m venv .venv

After installing the required Python packages

$ pip install -r requirements

you can now run the script

$ ./build_docs