6. Interactive CUDA C++ with Cling

The Cling CUDA extension brings the workflows of interactive C++ to GPUs, without losing performance and compatibility to existing software. Through this extension, C++ and CUDA can be used interactively on the target machine, allowing for optimization for particular models of accelerator hardware. The extension can be run on a Jupyter setup. Cling CUDA found application in the field of modeling of high-energy particles and radiation produced by high-energy laser facilities. In this framework, Cling CUDA allows for an interactive approach which enables relaunching only a wanted part of a simulation, starting from a given point which can be decided by the user.