NVIDIA announced at ISC 2024 that it will accelerate quantum computing efforts at national supercomputing centers worldwide with the open-source NVIDIA CUDA-Q™ platform. NVIDIA CUDA-Q simplifies hybrid code execution across diverse quantum processors, whether simulated or physical. Researchers can utilize cuQuantum-accelerated simulation backends or QPUs from partners or integrate their own simulator or quantum processor.
QPU Integration in Supercomputers with NVIDIA HPUs
Quantum Processing Unit integrated into supercomputers in Germany, Japan, and Poland running using NVIDIA-accelerated high-performance computing platforms.
- Germany’s Jülich Supercomputing Centre (JSC): JSC at Forschungszentrum Jülich will integrate a QPU from IQM Quantum Computers into its JUPITER supercomputer. The NVIDIA GH200 Grace Hopper Superchips will augment the JUPITER system to leverage quantum computing capabilities for faster calculations.
- Japan’s National Institute of Advanced Industrial Science and Technology: The ABCI-Q supercomputer at AIST forms part of Japan’s quantum computing initiative. It uses the NVIDIA Hopper architecture and will integrate a QPU from QuEra, increasing its quantum computing capability to drive research and innovation.
- Poznan Supercomputing and Networking Centre: PSNC has installed two photonic QPUs from ORCA Computing, which are connected to a new supercomputer partition accelerated by NVIDIA Hopper. This integration is meant to leverage quantum computing’s potential to make computations even more effective and faster in advancing science.
Advancing Quantum Computing Applications
At AIST, researchers will utilize Rubidium atoms controlled by laser light for quantum applications in AI, energy, and biology. These atoms act as qubits, similar to those in atomic clocks, offering the potential for a large-scale quantum processor.
PSNC’s quantum photonics systems use single photons for biology, chemistry, and machine learning research. These systems offer a scalable and modular architecture using standard telecom components.
JSC researchers will use superconducting qubits in JUPITER for chemical simulations and optimization. This setup aims to accelerate classical supercomputers with quantum computing.
Combining Quantum and Supercomputing
CUDA-Q firmly binds quantum computers to supercomputers, facilitating quantum computing alongside AI to address challenges like noisy qubits and optimize algorithms effectively.
An open-source and QPU-agnostic quantum-classical accelerated supercomputing platform, CUDA-Q is the preferred choice for most companies deploying QPUs, ensuring top-tier performance.
Features of CUDA-Q
- Easy Programming: User-friendly model extending C++ and Python for quantum and classical computing.
- GPU Support: Support for GPUs out of the box, which is useful for both preprocessing and postprocessing in quantum calculations, and for classical optimizations.
- Compiler Tools: Tools that compile quantum code into a format that machines understand for faster, more efficient operation.
- Performance: Initial benchmarks indicate a huge performance boost compared to the bare Python method, especially for larger systems.
- Library: Offers a set of basic quantum tools and functions to get users started with the language.
- Compatibility: Works on various types of quantum processors, whether they are physical or simulated, and integrates well with existing software written for CUDA.
Recommended: Top 20 AI Cloud Companies to Know in 2024
Benefits of NVIDIA CUDA-Q Platform for Quantum Computing
- Streamlines hybrid quantum-classical development with a unified programming model, improving productivity and scalability in quantum algorithm research.
- Connects to partner QPUs and GPU simulators, easy toolchain integration, and interoperates with modern GPU-accelerated applications.
- 2500X simulation speedup on a four A100 GPU for up to 26 qubits and scaling to 40 qubits by distributing the simulation across 128 GPU nodes.
Also Read: Top 10 CIO Influence News of Apr’24
FAQs
1. What are quantum processing units (QPUs)?
Quantum processing units (QPUs) are the brains of quantum computers, utilizing the behavior of particles like electrons or photons to perform calculations differently than traditional processors. They have the potential to speed up certain types of calculations significantly.
2. Which supercomputing sites are adopting CUDA-Q?
Supercomputing centers in Germany (Jülich Supercomputing Centre), Japan (National Institute of Advanced Industrial Science and Technology – AIST), and Poland (Poznan Supercomputing and Networking Center – PSNC) are integrating CUDA-Q into their systems.
3. How are QPUs being used at AIST, JSC, and PSNC?
At AIST, researchers are utilizing QPUs to explore quantum applications in AI, energy, and biology. JSC is developing quantum applications for chemical simulations and optimization problems, while PSNC is using QPUs to explore biology, chemistry, and machine learning.
4. What is the significance of integrating quantum with GPU supercomputing?
Tightly integrating quantum computers with GPU supercomputing enables accelerated quantum computing with AI, allowing researchers to solve complex problems more efficiently, such as dealing with noisy qubits and developing efficient algorithms.
5. What are the benefits of CUDA-Q?
CUDA-Q is an open-source and QPU-agnostic quantum-classical accelerated supercomputing platform. It delivers best-in-class performance and facilitates seamless integration and programming of multiple QPUs and GPUs, enabling innovative application areas in quantum-accelerated supercomputing.
[To share your insights with us as part of editorial or sponsored content, please write to sghosh@martechseries.com
