Quandela Publishes White Paper on Photonic QPU Integration With NVIDIA Infrastructure

Insider Brief

• Quandela and NVIDIA have published a technical white paper outlining an architecture for integrating Quandela’s photonic QPUs into existing AI and HPC infrastructure.
• The framework combines CPUs, GPUs and QPUs, with NVIDIA CUDA-Q, cuQuantum and NVQLink supporting quantum algorithm development, GPU simulation and low-latency QPU integration.
• The companies propose a three-step path covering access, integration and discovery, and scaling as organizations identify quantum workloads with potential value.

PRESS RELEASE — Building on our June announcement of a low-latency integration between a Quandela photonic QPU and NVIDIA AI infrastructure, the companies are publishing a technical white paper today. This paper outlines a vision and architectural framework for gradually integrating quantum computing into current AI and HPC environments.

This approach is based on the complementarity of CPUs, GPUs and QPUs. The CPU orchestrates workflows, while the GPU remains at the center of AI processing, notably handling intensive computation and simulation. The QPU acts as a specialized accelerator for workloads for which quantum computing could bring significant value.

Thanks to the NVIDIA CUDA-Q platform (an open quantum-GPU supercomputing programming platform) and NVIDIA cuQuantum (an SDK optimized for simulating quantum computations on GPUs), algorithms can be developed, simulated and tested on GPU before being run on an actual photonic QPU. NVIDIA NVQLink enables a low-latency connection between the NVIDIA GPU environment and the Quantum System Controller, which then controls the QPU.