Quantum version of CUDA, Nvidia releases revolutionary QODA programming platform
In 1999, NVIDIA defined the GPU, and the advent of the GPU is seen by the industry as the beginning of modern computer graphics technology. However, the GPU's microarchitecture is inherently suited for matrix-like parallel computing, and its capabilities are not limited to the graphics card domain, so from the early 21st century there were professional computationalists who wanted to use the GPU to do some parallel computing related to the artificial intelligence domain. However, a lot of underlying language code had to be written to invoke the computational power of the GPU.
Simply put, the GPU was too big for graphics only, so NVIDIA introduced CUDA (Unified Computing Architecture) in 2006, which enabled GPUs to use their parallel computing power easily and efficiently, making them more than just graphics cards and becoming general-purpose processors. CUDA has now become the central node for connecting AI, and CUDA+GPU system has greatly promoted the development of AI field.
Now, there is a new opportunity in front of NVIDIA - hybrid quantum-classical computing. In order to play a role in the early days of quantum computing, hybrid quantum-classical computing was created. The advantage of hybrid computing is that most of the tedious tasks are still left to classical computing, while quantum computing only handles the tasks it is good at.
However, there is no unified computing platform that combines quantum computing (QPU, quantum processing unit) with classical computing (CPU and GPU). For this reason, on July 12, 2022, NVIDIA released a quantum version of its unified computing platform, QODA (Quantum Optimized Device Architecture), for accelerating breakthroughs in quantum R&D in artificial intelligence, high-performance computing, health, finance, and other disciplines.
QODA will be applied in many fields
NVIDIA actually sees quantum computing as another element of the heterogeneous high-performance computing (HPC) system architecture and envisions a programming model that seamlessly integrates quantum co-processing into its existing CUDA ecosystem.
QODA aims to make quantum computing more accessible by creating a coherent hybrid quantum-classical programming model.QODA is an open, unified environment for some of today's most powerful computers and quantum processors (QPUs) that will increase scientific productivity and enable quantum research at a larger scale.
NVIDIA says QODA is inherently interoperable with existing classical parallel programming models such as CUDA, OpenMP and OpenACC. This compiler implementation also reduces the quantum-classical C++ source code representation to a binary executable that natively supports cuQuantum's (NVIDIA's quantum acceleration tool) simulation backend as a target.
This programming and compilation workflow enables a well-performing programming environment through standard interoperability with GPU processing and circuit emulation to accelerate hybrid algorithm development activities that can scale from laptops to distributed multi-node, multi-GPU architectures.
With QODA, HPC and AI domain experts can easily add quantum computing to existing applications, leveraging today's quantum processors, as well as simulated future quantum machines using NVIDIA DGX systems and the vast array of NVIDIA GPUs available in scientific supercomputing centers and public clouds.
Tim Costa, Product Lead for High Performance Computing and Quantum Computing at NVIDIA, said, "A hybrid solution that combines classical and quantum computing will enable scientific breakthroughs in the short term. qODA will revolutionize quantum computing by providing developers with a powerful and efficient programming model."
The advantages of QODA are as follows.
Flexible and scalable: supports hybrid deployments by emulating on a single GPU, the NVIDIA DGX SuperPOD™ supercomputer and multiple QPU partner backends.
Open: connection to any type of QPU backend, allowing access by all users.
High performance: end-to-end Variational Quantum Element Solver (VQE) with 20 quantum bits based on the QODA framework has 287 times higher performance and significantly higher scalability compared to the Pythonic framework
Easy integration: interoperate with modern GPU-accelerated applications
Efficient: simplifies hybrid quantum-classical development with a unified environment to improve productivity and scalability of quantum algorithm research.
QODA features the following：
Extending C++'s core-based programming model for hybrid quantum-classical systems (full Python support coming soon)
Native support for GPU hybrid computing, with support for GPU pre-processing and post-processing and classical optimization
System-level compiler toolchain with NVQ++ compiler for quantum cores split and compiled down to multi-level intermediate representation (MLIR) and quantum intermediate representation (QIR)
Initial NVQ++ benchmarks show a 287x improvement in end-to-end VQE performance for 20 quantum bits compared to standard Pythonic implementations, with significant scalability improvements as system size increases
Standard library of quantum algorithm primitives
Interoperate with partner QPUs as well as simulated QPUs using the cuQuantum GPU platform; work with QPU builders of many different quantum bit types
Leading quantum organizations are already using NVIDIA GPUs and the highly specialized NVIDIA cuQuantum to develop individual quantum circuits, NVIDIA said. With QODA, developers can build complete quantum applications simulated with NVIDIA cuQuantum on a GPU-accelerated supercomputer.
At the Q2B conference in Tokyo on July 12, NVIDIA announced QODA's collaboration with quantum hardware providers IQM Quantum Computers, Pasqal, Quantinuum, Quantum Brilliance and Xanadu; software providers include QC Ware and Zapata Computing. Supercomputing centers include the Jülich Center in Germany, Lawrence Berkeley National Laboratory and Oak Ridge National Laboratory.
QODA's Partners
Quantinuum (formerly Honeywell) is working with NVIDIA to enable users of Quantinumum H-Series quantum processors powered by Honeywell to program and develop next-generation hybrid quantum-classical with QODA," said Alex Chernoguzov, chief engineer at Quantinuum (formerly Honeywell). applications. This links the best performing classical computers with our world-class quantum processors."
The hybrid quantum-classical capabilities developed by NVIDIA will enable HPC developers to accelerate their existing applications by providing an efficient way to program quantum and classical resources in an integrated environment," said Yudong Cao, chief technology officer at Zapata. As a result, near-term applications in chemistry, drug discovery, materials science and other fields can now be seamlessly integrated with quantum computing and drive new discoveries in these fields as practical quantum advantages emerge."
The platform is still under development and is expected to be available for beta users in late 2022, with full availability in early 2023.
[1]https://developer.nvidia.com/qoda
[2]https://nvidianews.nvidia.com/news/nvidia-announces-hybrid-quantum-classical-computing-platform
