Adding CUDA® Support to Cling: JIT Compile to GPUs

We present the results of a diploma thesis adding CUDA (runtime) C++ support to cling. Today's HPC systems are heterogeneous and get most of their computing power from so-called accelerator hardware, such as GPUs. Programming GPUs with modern C++ is a perfect match, allowing perfectly tailored and zero-overhead abstractions for performance-critical "kernels".

Nevertheless, tool complexity in development and debugging can be discouraging for new users. We are addressing this by not only adding low-level support for accelerators but also by going up the open source software-stack enabling interactive, CUDA C++ Jupyter notebooks, e.g. through xeus-cling.