NVIDIA and Google Partner to Silence Quantum Noise with Cutting-Edge Simulations

NVIDIA and Google have joined forces to accelerate the design of Google's next-generation quantum computing processors. On Monday, November 18, NVIDIA announced its collaboration with Google Quantum AI, the team dedicated to developing quantum computing hardware and software tools. This partnership leverages simulations supported by NVIDIA’s CUDA-Q platform to advance the development of Google’s upcoming quantum devices.

Google Quantum AI is utilizing a hybrid quantum-classical computing platform and NVIDIA’s Eos supercomputer to simulate the physical properties of its quantum processors. This collaborative effort aims to overcome the current limitations of quantum computing hardware, primarily attributed to what researchers refer to as "noise." Currently, quantum computing devices can only execute a limited number of quantum operations before noise forces computations to a halt.

Noise poses a significant challenge for quantum processors, which are extremely sensitive to disturbances like stray photons generated by heat, random signals from nearby electronic devices, and physical vibrations. These interferences can rapidly destroy the delicate quantum states known as superpositions, substantially affecting the accuracy of quantum computations. Progress in building practical quantum computers, therefore, hinges on the ability to effectively "tame" noise.

Guifre Vidal, a research scientist with Google Quantum AI, emphasized the importance of controlling noise while scaling up quantum hardware. He stated, "We need to control the noise as we expand the scale of quantum hardware to develop commercially viable quantum computers. With NVIDIA's accelerated computing, we are exploring the impact of larger quantum chip designs on noise."

The endeavor to understand noise in quantum hardware design requires intricate dynamical simulations that can fully capture how qubits interact with their surroundings. These simulations are usually computationally expensive. However, NVIDIA claims that Google can now conduct one of the largest and fastest dynamic simulations of quantum devices globally at a fraction of the cost by using the CUDA-Q platform on its Eos supercomputer, employing 1,024 H100 Tensor Core GPUs.

Thanks to NVIDIA’s CUDA-Q and H100 GPUs, Google can perform comprehensive and realistic simulations on devices containing 40 qubits—a scale never achieved in similar simulations before. The simulation technology provided by CUDA-Q reduces what used to take a week into a matter of minutes.

NVIDIA has publicly provided software on its CUDA-Q platform to support these accelerated dynamic simulations, enabling quantum hardware engineers to quickly scale their system designs.

Tim Costa, NVIDIA’s Director of Quantum and HPC, remarked, "AI supercomputing capabilities will contribute to the successful advancement of quantum computing. Google’s use of the CUDA-Q platform demonstrates the pivotal role of GPU-accelerated simulations in helping solve real-world problems through quantum computing."