NVIDIA and Google Collaborate to Enhance Quantum AI Processor Design
NVIDIA collaborates with Google Quantum AI to enhance quantum computing device design through simulations. The partnership aims to address noise limitations in quantum computing hardware. Utilising NVIDIA's CUDA-Q platform and H100 GPUs, Google conducts extensive simulations of devices with 40 qubits.
This collaboration involves leveraging simulations powered by the NVIDIA CUDA-Q™ platform to enhance the design process.
Google Quantum AI is utilising a hybrid quantum-classical computing platform alongside the NVIDIA Eos supercomputer to simulate the physics of its quantum processors. This strategic move aims to address the existing limitations of quantum computing hardware, particularly in managing noise that can disrupt quantum operations.
Guifre Vidal, a research scientist at Google Quantum AI, emphasised the significance of controlling noise for the practical application of quantum computers. By harnessing NVIDIA's accelerated computing capabilities, the team is delving into the noise implications associated with scaling up quantum chip designs.
The complexity of understanding noise in quantum hardware necessitates intricate dynamical simulations that accurately depict how qubits interact within a quantum processor and with their surroundings. Historically, these simulations have been financially burdensome to pursue due to their computational intensity.
However, with the CUDA-Q platform, Google can leverage 1,024 NVIDIA H100 Tensor Core GPUs within the NVIDIA Eos supercomputer to conduct one of the world's most extensive and rapid dynamical simulations of quantum devices at a reduced cost.
Tim Costa, NVIDIA's director of quantum and HPC, highlighted the pivotal role of AI supercomputing power in the success of quantum computing. He commended Google's adoption of the CUDA-Q platform, underscoring the importance of GPU-accelerated simulations in advancing quantum computing to address real-world challenges.
By utilising CUDA-Q and H100 GPUs, Google can execute comprehensive and realistic simulations of devices containing 40 qubits, marking a significant milestone in simulation capabilities. The accelerated simulation techniques facilitated by CUDA-Q have drastically reduced the time required for noisy simulations, enabling tasks that previously took a week to be completed within minutes.
The software underpinning these accelerated dynamic simulations will be made publicly accessible through the CUDA-Q platform, empowering quantum hardware engineers to efficiently scale their system designs.
NVIDIA collaborates with Google Quantum AI to enhance quantum computing device design through simulations.
The partnership aims to address noise limitations in quantum computing hardware.
Utilising NVIDIA's CUDA-Q platform and H100 GPUs, Google conducts extensive simulations of devices with 40 qubits.
Source: NVIDIA