At SC24, NVIDIA announced its collaboration with Google Quantum AI to advance the design of quantum processors through large-scale simulations on the NVIDIA CUDA-Q™ platform. Using NVIDIA Eos supercomputers powered by 1,024 H100 Tensor Core GPUs, Google Quantum AI is simulating quantum device physics to address hardware limitations caused by noise in quantum operations.
These simulations focus on understanding how qubits interact with their environment, a complex process requiring significant computational resources. The CUDA-Q platform enables simulations of up to 40-qubit devices—among the largest ever performed—dramatically reducing computation times from weeks to minutes. This breakthrough aids in scaling quantum hardware designs while mitigating noise.
The collaboration highlights the role of GPU-accelerated AI supercomputing in quantum research, with Google using these tools to model increasingly complex quantum processors. NVIDIA plans to make the software powering these simulations publicly available through CUDA-Q, empowering researchers globally to optimize quantum hardware development.
This partnership underlines the synergy between classical and quantum computing in tackling real-world challenges and advancing quantum technologies. For more details, visit NVIDIA’s press release here.
November 20, 2024
