How GPUs, QPUs, and CPUs Are Unite by NVIDIA CUDA-Q
NVIDIA announced that the open-source NVIDIA CUDA-Q technology will accelerate quantum computing efforts at national supercomputing centres across the globe.Advance Quantum Computing Research by Including Grace-Hopper and Quantum-Classical Accelerated Supercomputing Platform Supercomputers in Germany, Japan, and Poland
CUDA-Q from NVIDIA
Innovative hybrid quantum-classical computing system
For algorithm research and quantum advantage applications, bridging technology is needed to enable dynamic workflows across system architectures. Using a single, open programming model, NVIDIA CUDA-Q is an open-source platform that integrates and programmes GPUs, CPUs, and quantum processing units (QPUs) in a single system. NVIDIA CUDA-Q enables GPU-accelerated system performance and scalability across heterogeneous QPU, CPU, GPU, and simulated quantum system parts.NVIDIA CUDA-Q offers a single programming model designed for a hybrid environment where CPUs, GPUs, and QPUs work together to make the development of hybrid applications easier. It consists of a system-level toolchain enabling Python and C++ language extensions and application acceleration.
Main Benefits
Productive
improves the efficiency and scalability of quantum algorithm development by making it easier to create hybrid quantum-classical systems with a single programming model.Flexible Organisation
Connects to partner QPUs and GPU simulators, readily integrates with toolchains, and interfaces with modern GPU-accelerated programmes.Outstanding Results
A 2500X speedup on four A100 GPUs can simulate increase to 26 qubits, and distributing the simulation over 128 GPU nodes can simulate 40 qubits.Supercomputing centres in Germany, Japan, and Poland will use the platform to power their quantum processing units (QPUs) in NVIDIA-accelerated high-performance computing systems.
The brains of quantum computers are known as quantum processor units, or QPUs. Through the use of particle behaviour in their calculations, such as that of electrons or photons, they may be able to do some computations faster than traditional computers.
The Jülich Supercomputing Centre (JSC) at Forschungszentrum Jülich will use IQM Quantum Computers' QPU in addition to the JUPITER supercomputer, which is driven by the NVIDIA GH200 Grace Hopper Superchip.
The ABCI-Q supercomputer is located at Japan's National Institute of Advanced Industrial Science and Technology (AIST) and is meant to advance the nation's quantum computing initiative. The machine will be powered by the NVIDIA Hopper design and equipped with a QuEra QPU.
Two ORCA Computing photonic QPUs have been integrated by the Poznan Supercomputing and Networking Centre (PSNC) in Poland. These QPUs are connected to a recently created NVIDIA Hopper-powered supercomputer partition.
"Tight integration of quantum with GPU supercomputing will enable useful quantum computing," said Tim Costa, president of NVIDIA's quantum and HPC division. NVIDIA's quantum computing platform enables innovators like as AIST, JSC, and PSNC to advance the state of the art in quantum-integrated supercomputing and push the boundaries of scientific discovery.
Thanks to the integration of QPU with ABCI-Q, researchers at AIST will be able to investigate quantum applications in AI, energy, and biology by employing laser-controlled Rubidium atoms as qubits to execute calculations. These atoms are exactly the same as those used in accurate atomic clocks. Since all atoms are the same, this presents a viable approach to creating a large-scale, high-fidelity quantum processor.
"Japan's researchers will make progress towards practical quantum computing applications with the ABCI-Q quantum-classical accelerated supercomputer," said Masahiro Horibe, deputy director of G-QuAT/AIST. "With support from NVIDIA, these trailblazers are expanding the frontiers of quantum computing research."
PSNC's QPUs will enable researchers to explore biology, chemistry, and machine learning with two PT-1 quantum photonics devices. The devices use single photons, or light packets at telecom frequencies, as qubits. This enables the creation of a distributed, scalable, and modular quantum infrastructure using widely accessible, ordinary telecom components.
"Our partnership with ORCA and NVIDIA has allowed us to create a unique environment and build a new quantum-classical hybrid system at PSNC," said Krzysztof Kurowski, CTO and deputy director of PSNC. Multiple QPUs and GPUs that are handled by user-centric services must be openly deployed and programmed for by developers and users. This strong collaboration paves the way for a new generation of quantum-accelerated supercomputers for many cutting-edge application domains.
Because of the QPU's connection with JUPITER, researchers at JSC will be able to develop quantum applications for chemical simulations and optimisation problems. They will also be able to demonstrate how classical supercomputers can be sped up by quantum computers. The fundamental components of this technology are electrical resonant circuits, or superconducting qubits, which can be made to operate at low temperatures similar to artificial atoms.
Kristel Michielsen, head of JSC's quantum information processing division, said, "Quantum computing is getting closer with hybrid quantum-classical accelerated supercomputing." "Through our ongoing partnership with NVIDIA, JSC researchers will advance the disciplines of quantum computing, chemistry, and material science.”
By closely integrating quantum computers with supercomputers, CUDA-Q enables quantum computing with AI to address problems such as noisy qubits and provide efficient algorithms.
CUDA-Q is a GPU-independent, open-source quantum-classical accelerated supercomputing platform. QPUs are used by most enterprises because they provide best-in-class performance.
News Source : QPUs
0 Comments