NVIDIA expands quantum computing horizons with AI supercomputing

NVIDIA is leading the advancement of quantum computing by integrating AI supercomputing capabilities, as recently announced at Supercomputing 2024 (SC24). According to NVIDIA, the tech giant has partnered with a variety of industry and academic leaders to address the current challenges of quantum computing and push the boundaries of the rapidly evolving technology.

AI meets quantum computing

Generative AI is playing a critical role in overcoming quantum computing challenges. NVIDIA collaborated with scientists to publish a research paper titled “Artificial Intelligence for Quantum Computing,” exploring the transformative potential of AI in this field. This paper highlights the use of AI models such as GPT to synthesize quantum circuits and decode quantum error correction codes, demonstrating significant progress at the intersection of these two innovative fields.

Innovative infrastructure and algorithms

The collaboration between Poznan Supercomputing and Networking Center (PSNC), ORCA Computing, and NVIDIA has resulted in the first fully featured multi-QPU, multi-GPU, multi-user infrastructure. This setup leverages NVIDIA H100 Tensor Core GPUs and the NVIDIA CUDA-Q platform to showcase a new resource state generator with a pre-trained transformer for photonic quantum processors (RS-GPT) algorithm. The collaboration also developed a hybrid quantum-classical generative adversarial network (GAN) for facial recognition and a quantum neural network for medical diagnosis.

Integration of quantum and classical technologies

There were several announcements at SC24 highlighting CUDA-Q’s integration with a variety of quantum hardware providers, including Anyon, Fermioniq, and QuEra. These partnerships expand access to accelerated quantum supercomputing by seamlessly integrating quantum resources into hybrid algorithms.

Enhancing quantum hardware design

NVIDIA’s CUDA-Q platform now supports quantum hardware development using AI supercomputing. New dynamic simulation capabilities in CUDA-Q 0.9 enable high-precision simulation of quantum systems, helping QPU vendors improve their hardware designs and qubits. NVIDIA is also collaborating with Google Quantum AI to perform large-scale simulations of transmon qubits to advance our understanding of quantum device physics.

Extensive access to advanced algorithms

CUDA-QX, a collection of application-specific libraries optimized for GPU acceleration, provides researchers with the tools to explore the next generation of quantum computing topics. These libraries include quantum error correction code and solvers for applications such as chemical simulations and leverage GPU supercomputing to provide unprecedented scalability.

Collaborative Innovation

NVIDIA is actively collaborating with institutions such as Yale University and Moderna to explore quantum transducers for molecule creation and quantum neural networks for predicting biomolecular binding affinity. Our collaboration with Hewlett Packard Enterprise focuses on improving circuit knitting technology, and our partnership with Algorithmiq accelerates noise mitigation in quantum computing.

Through these partnerships and initiatives, NVIDIA is making significant contributions to the quantum computing ecosystem by providing tools and platforms that integrate AI and quantum technologies to solve complex computing problems.

Image source: Shutterstock