Impactful quantum applications are hybrid in nature, meaning they take advantage of both classical and quantum resources for computing. While the integration of quantum processors (QPUs) into HPC systems gives rise to unique opportunities for accelerating subroutines that would be prohibitive to execute on classical processors, it also poses unique challenges for coordination between different processors.

Integration efforts today are still in their infancy, and further research is needed to explore how to build and leverage heterogenous quantum-classical systems to their fullest extend. While QPUs continue to grow in size and capabilities, it is important to co-develop the classical hardware and software needed for efficient compilation and execution of hybrid applications.

CUDA-Q is an open-source platform for quantum computing that offers a unified programming model designed for CPUs, GPUs, and QPUs to work together effectively in a high-performance system. In this talk, I will introduce the CUDAQ-toolchain and how to take advantage of GPUs for accelerating the development and execution of quantum applications today. I will discuss current integration strategies before outlining some of the necessary steps to allow for more tightly coupled interactions between classical and quantum resources in future large-scale systems.