Nvidia aims to bridge GPU and quantum computing fields via cuQuantum

Nvidia Accelerates Efforts to Bridge GPU and Quantum Computing Fields with Tensor-capable cuQuantum (opens in a new tab)quantum simulation toolbox. Through this, the company aims to accelerate quantum circuit simulation workloads in a way beyond the reach of current Noisy Intermediate-Scale Quantum (NISQ) systems. But for this, the company is betting on further integration between quantum and classical systems towards hybrid solutions. Unsurprisingly, GPUs are at the forefront of Nvidia’s quantum developments.

Nvidia cuQuantum

(Image credit: Nvidia)

Nvidia aims to create a low-latency connection that can bridge its GPUs — and their quantum simulation-capable Tensor Cores — with current and future Quantum Processing Units (QPUs). The goal here is to take advantage of the extremely powerful parallel processing of the GPU, harnessing them for quantum-specific workloads such as circuit optimization, calibration, and error correction, while removing the bottleneck bottleneck of communications between quantum and classical systems.

Another element of Nvidia’s approach to quantum computing aims to offer a common software layer not unlike the company’s CUDA programming model. (opens in a new tab).

The idea is that this programming model greatly simplifies code-level interaction with QPUs and quantum simulations, which is still done in what amounts to low-level assembly code. The goal is to streamline a unified, quantum programming model and compiler toolchain (opens in a new tab) which summarizes different QPUs for a more targeted use of quantum capabilities. Nvidia hopes to ease the transition from classical to classical quantum workloads by allowing users to partially port their High-Performance Computing (HPC) applications to a simulated QPU and then to the processor itself.

According to NVidia, dozens of organizations are already leveraging its cuQuantum toolkit to support their quantum work. Amazon Web Services already offers cuQuantum integration through its Braket service (opens in a new tab), exhibiting a 900x speedup on quantum machine learning workloads. Other platforms leveraging Nvidia’s cuQuantum include Google’s Qsim, IBM’s Qiskit Aer, Xanadu’s PennyLane, Classiq’s quantum algorithm design platform. Nvidia recently broke a world record (opens in a new tab) in quantum computing simulation by leveraging its cuQuantum framework and ultra-powerful Selene supercomputer, powered by its SuperPOD DGX (opens in a new tab).

Joining Nvidia’s developing cuQuantum ecosystem is Menten AI, a drug discovery startup that aims to leverage cuQuantum’s tensor network library to simulate protein interactions and novel drug molecules. The goal is to accelerate drug design, whose workloads are naturally suited to the probabilistic nature of quantum computing.

“While quantum computing hardware capable of running these algorithms is still under development, classical computing tools such as NVIDIA cuQuantum are crucial to advancing the development of quantum algorithms,” said Alexey Galda, Principal Scientist at Menten. HAVE.

Nvidia has achieved remarkable penetration into the HPC market through its CUDA software stack, and it looks like the company is aiming to repeat the feat for the quantum realm via cuQuantum. In what is one of the most complex research fields in the world, it certainly seems like a streamlined software package would help accelerate the road to quantum by leaps and bounds.

Sherry J. Basler