NVIDIA Makes Waves in the Prep for Quantum Computing

The space of quantum computing is brimming with potential and possibilities—and NVIDIA, powered by its Blackwell architecture, has an array of quantum computing workloads in development. As the developers of accelerated computing’s leading architectures—NVIDIA GB200 NVL72 systems and their fifth-generation multi node NVIDIA NVLink interconnect capabilities—NVIDIA is paving the way in today’s development of tomorrow’s quantum tech, according to the company.

First of NVIDIA’s initiatives is developing better quantum algorithms by simulating candidate algorithms and collecting data that will help improve performant quantum applications. While these simulations are traditionally computationally intensive, GB200 NVL72’s high-bandwidth interconnect with all-to-all GPU connectivity enables NVIDIA to execute these simulations on feasible time scales—with an 800x speedup compared with the best CPU implementations, the company noted.

NVIDIA is redefining chip manufacturing, relying on the previously mentioned simulations to create more performant processor designs. These low-noise qubit designs—discovered and refined through NVIDIA simulations accelerated by GB200 NVL72 and cuQuantum’s dynamics library—are crucial for quantum computing’s future.

While AI models are showing some promise for optimizing quantum computing, they run into a common challenge: obtaining the vast repositories of data needed to train them successfully. According to simulated quantum processors, GB200 NVL72 can output quantum training data 4,000x faster compared to CPU-based techniques, enabling quantum computing to benefit from AI’s continuous advancements.

According to Timothy Costa, senior director of CAE, quantum and CUDA-X at NVIDIA, “Effective future quantum applications will lean on both quantum and classical hardware, seamlessly distributing algorithm subroutines to whichever hardware type is most appropriate.” Meaning, hybrid algorithms will be necessary to accommodate these environments, combining both simulations of quantum hardware with access to state-of-the-art AI supercomputing.

NVIDIA CUDA-Q, the QPU-agnostic platform for accelerated quantum supercomputing, enables such hybridity, offering an ideal hybrid computing environment that allows researchers to explore hybrid quantum-classical applications—and speeds development by 1,300x, according to NVIDIA.

Finally, GB200 NVL72 conquers the challenges of quantum error correction—a control process that continually processes terabytes of qubit data through demanding decoding algorithms—with a 500x acceleration in running a commonly used class of decoding algorithms.

“These breakthroughs are allowing the quantum computing industry to perform the quantum-GPU integrations needed for large-scale useful quantum computing,” said Costa. “NVIDIA is working toward a future where all supercomputers integrate quantum hardware to solve commercially relevant problems. NVIDIA GB200 NVL72 is the platform for building this future.”

To learn more about NVIDIA’s latest quantum computing advancements, please visit https://www.nvidia.com/en-us/.