Quandela is collaborating with NVIDIA on the development of NVIDIA NVQLink™, a new open architecture bringing NVIDIA accelerated computing to the quantum domain.
Building on NVIDIA’s expertise in accelerated computing and interconnect technologies, NVQLink, integrated with NVIDIA CUDA-Q, introduces a model where quantum processing units (QPUs) can be tightly coupled with a GPU-based control system, forming a “logical QPU”. This design enables real-time, deterministic data exchange between HPC resources and quantum systems.
NVIDIA NVQLink™ is the open platform needed to build logical qubits. It allows QPU builders and application developers to perform real-time orchestration of QPUs and their control hardware alongside CPUs and GPUs. Through NVIDIA CUDA-Q, NVQLink exposes a low-latency, high-throughput API that enables developers to seamlessly interlace classical and quantum compute resources. NVQLink also acts as an interface for data transfer between logical qubit systems and conventional AI supercomputing environments, supporting hybrid quantum-classical applications. The platform is designed with maximal flexibility, accommodating heterogeneous quantum hardware control architectures.
Quandela brings NVQLink to photonic QPUs, integrating them within this GPU-centric ecosystem.
Our collaboration aims to:
- Enhance reliability and performance of quantum computation through continuous calibration, real-time error mitigation, and AI-driven surrogate models accelerated on GPUs.
- Accelerate hybrid AI-Quantum workflows, leveraging low-latency communication between classical GPU compute nodes and quantum controllers – as envisioned in our MerLin framework.
- Enable tight integration within HPC infrastructures, introducing quantum and other unconventional computing models into the core of modern supercomputing.
- Foster interoperability across modalities – photonic, superconducting, trapped ions, and beyond – through a unified software and runtime stack.
“NVQLink provides the low-latency interconnect needed to introduce quantum hardware into modern supercomputing”, said Tim Costa, General Manager for Quantum at NVIDIA, “Quandela is scaling photonic quantum processing, and their collaboration on NVQLink provides them access to the accelerated computing platform for controlling qubits and developing hybrid applications”
Quandela, with NVIDIA, is building the bridge between AI supercomputing and real-time quantum control, paving the way for scalable, reliable, and hybrid quantum–classical systems.
What NVQLink Brings to Quandela Users ?
For Quandela users, the integration of NVIDIA NVQLink™ means faster, more reliable access to photonic QPUs. Through real-time orchestration between the GPU control layer and the quantum processor, users will benefit from:
- Improved sampling quality, thanks to on-the-fly circuit transpilation and real-time error mitigation.
- Continuous, dynamic calibration, ensuring optimal system performance during live experiments.
- Extended compilation capabilities, leveraging GPU acceleration for advanced circuit decomposition and hybrid algorithm optimization.
- Seamless hybrid execution, where GPUs and QPUs communicate in real time to support AI-enhanced quantum workflows, as pioneered in Quandela’s MerLin framework.
This collaboration enhances the user experience with more stable, adaptive, and intelligent photonic quantum computing, bridging classical AI supercomputing and quantum hardware.
Read more in NVIDIA’s announcement and white paper.
