Nvidia’s new platform could usher in the era of hybrid quantum-classical computing

Quantum computing is on the horizon, potentially bringing with it huge advances in areas like pharmaceuticals, electric vehicle engineering, investing, and more. 

Realistically, though, most of the applications used to power these kinds of advances won’t rely solely on quantum computing. They’ll likely use hybrid quantum-classical computing, with some algorithms optimized for quantum computing and others running on classical computing infrastructure. 

Currently, though, there’s a missing piece of the puzzle: a way for most high-performance computing (HPC) developers to accelerate their existing applications with quantum computers. 

SEE: What is quantum computing? Everything you need to know about the strange world of quantum computers

To overcome this problem, Nvidia is introducing a new platform called QODA — Quantum Optimized Device Architecture — that will create a unified, open environment across traditional computers and quantum processors. 

Nvidia contends that, with open and interoperable standards, developers can use across different hardware and software platforms and hybrid quantum-classical computing will be within reach.

Tim Costa, Nvidia’s director of HPC and quantum computing products, said the platform will prompt a “revolution in the accessibility of hybrid quantum-classical computing”. 

QODA includes a programming model and a compiler toolchain that enables quantum acceleration of today’s scientific computing applications. This means the programming model and toolchain are built to integrate into and interoperate with the tools, the programming models, and the languages that today’s scientific developers are already using in their applications.

Nvidia expects to have a closed beta of QODA up and running by the end of the year. When it’s available, the platform will be free. 

Nvidia doesn’t build quantum processors — it builds the classical GPU supercomputing infrastructure that will work in tandem with quantum processors in a hybrid setup. That’s why the QODA platform is an open platform. Nvidia is building it in collaboration with the leading quantum computing hardware companies and software companies, as well as research institutions. 

At launch, Nvidia will have five leading quantum computing hardware providers across five different qubit modalities that will all be supported by QODA, including IQM Quantum Computers, Pasqal, Quantinuum, Quantum Brilliance, and Xanadu. Its launch partners also include software companies QC Ware and Zapata, as well as the supercomputing centers at Forschungszentrum Jülich, Lawrence Berkeley National Laboratory, and Oak Ridge National Laboratory.

Costa likened Nvidia’s development of QODA to the company’s development of CUDA — the parallel computing platform and programming model that allows developers to accelerate their compute-intensive applications with GPUs. 

“It’s really part of our DNA as a company to improve accessibility of disruptive technologies,” Costa said to ZDNet. “We did it with GPU computing, and now we’re looking at the same problem in hybrid quantum-classical computing.”