Nvidia has made clear its intention to be a full computing stack provider, and it is becoming increasingly clear this ambition extends to the emerging sector of quantum computing, where the company this week announced a new unified programming layer that will help developers to more quickly ready their applications for hybrid classical-quantum computing systems.
This announcement is the latest in a series of steps the company has made to move into quantum computing, over the last two years.
Much of the most valuable quantum computing work being done today is being done in this hybrid manner where users are looking to align and coordinate classical GPU computing resources and new quantum processing unit resources to create simulations and emulations that would be difficult or impossible to accomplish on classical systems alone. Nvidia’s new Quantum Optimized Device Architecture, announced this week at a quantum computing conference in Tokyo, allows developers to leverage a framework of code, language and cues that is more familiar and easier to work with than the assembly code many quantum computers use.
“Quantum computers today are programmed in the equivalent of assembly code, which has an extremely steep learning curve for researchers who are not already quantum engineers incorporating quantum computing into their workflows,” said Tim Costa, director of HPC and Quantum Computing Products at Nvidia, in an email to IQT News. “QODA removes that barrier to entry by enabling programming of hybrid quantum systems in a model familiar to scientific computing developers, interoperable with the best of today’s classical computing applications… This opens the door to domain scientists working in all application areas to develop quantum algorithms and integrate them directly into their applications.”
Costa compared QODA’ to Nviidia’s CUDA, the software framework that the company released more than a decade ago to help developers accelerate their work with Nvidia GPUs.
“In the same way CUDA revolutionized scientific developer access to the GPU, QODA revolutionizes developer access to hybrid quantum-classical computing by providing a familiar programming environment that interoperates with today’s applications, programming models and software stacks,” he said. “The environment is uniform across all quantum computing resources, emulated and physical.”
Costa added, “A common platform is necessary for the development of hybrid quantum-classical computing applications. QODA provides a unified framework for this development that, in addition to being familiar to domain scientists and interoperable with today’s scientific computing applications, is also a single framework that supports any quantum computing resource.”
QODA can work with Nvidia’s cuQuantum software developers kit, its ncq++ programming language and its DGX systems, but to Costa’s point it also is being made available to Nvidia partners. :It’s open to partners, and in addition to having it run on a cuQuantum-accelerated emulated quantum resource, we’re working with five quantum hardware companies (IQM, Pasqal, Quantinuum, Quantum Brilliance, and Xanadu) representing five qubit modalities (Superconducting, Neutral Atom, Ion Trap, Diamond, and Photonics), to support their processors in the initial alpha release, and will add many more,” Costa said.
Kristel Michielsen, who manages the Jülich near Cologne, Germany, said in a blog posted by Nvidia, “This helps bring quantum computing closer to the HPC and AI communities,” she said. “It will speed up how they get things done without them needing to do all the low-level programming, so it makes their life much easier.”
Marcus Doherty, Co-Founder and Chief Scientific Officer, Quantum Brilliance, also commented, saying in a statement, “Nvidia revolutionized high performance computing, and its new QODA platform is a bold step forward in innovating the quantum industry as well. Our unique room-temperature diamond quantum microprocessor exploits this hybrid approach and will add a critical new element to the HPC and embedded computing landscape. It is essential that near-term quantum computing requires coupling of classical and quantum hardware to realize the technology’s potential. This allows sources of classical computing power such as an Nvidia graphics processing unit (GPU) to be much closer to the quantum processing unit (QPU). The Nvidia QODA platform provides the unifying framework for this vision.”
In past interviews, Nvidia's Costa has suggested that Nvidia's growing efforts in the quantum space should allow it to be identified as a quantum computing company. So far, those efforts have not included developing its own QPU in the same way it has developed GPUs, CPUs and DPUs, and Costa has said the company has no current plans to announce a QPU, but to be a true full stack provider in this emerging sector, that could be necessary.