Nvidia today announced that it can speed up quantum computing efforts at national supercomputing centers world wide with the open source Nvidia CUDA-Q platform.
Supercomputing sites in Germany, Japan and Poland will use the platform to power the quantum processing units (QPUs) of their Nvidia-accelerated high-performance computing systems. Nvidia also announced that nine recent supercomputers worldwide are using Nvidia Grace Hopper superchips to speed up scientific research and discovery. Together, the systems deliver 200 exaflops, or 200 trillion calculations per second, of energy-efficient AI computing power.
QPUs are the brains of quantum computers that use the behavior of particles akin to electrons or photons to calculate in another way than traditional processors, with the potential to perform certain forms of calculations faster.
The Jülich Supercomputing Center (JSC) at Forschungszentrum Jülich (FZJ) is installing a QPU built by IQM Quantum Computers to enhance its Jupiter supercomputer, which is supplied with the Nvidia GH200 Grace Hopper superchip.
The ABCI-Q supercomputer on the National Institute of Advanced Industrial Science and Technology (AIST) in Japan is designed to advance the country's quantum computing initiative. The system relies on the Nvidia Hopper architecture and can add a QPU from QuEra.
Poland's Poznan Supercomputing and Networking Center (PSNC) recently installed two photonic QPUs built by ORCA Computing connected to a brand new supercomputer partition accelerated by Nvidia Hopper.
“Useful quantum computing is enabled by the tight integration of quantum and GPU supercomputing,” said Tim Costa, director of quantum and HPC at Nvidia, in an announcement. “Nvidia’s quantum computing platform enables pioneers like AIST, JSC and PSNC to push the boundaries of scientific discovery and advance the state-of-the-art in quantum integrated supercomputing.”
The QPU integrated into ABCI-Q will enable researchers at AIST to explore quantum applications in AI, energy and biology, using rubidium atoms controlled by laser light as qubits to perform calculations. These are the identical kind of atoms which might be utilized in precision atomic clocks. Each atom is an identical, providing a promising method for producing a large-scale, high-precision quantum processor.
“Japan's researchers will make progress toward practical quantum computing applications with the quantum classical accelerated supercomputer ABCI-Q,” Masahiro Horibe, deputy director of G-QuAT/AIST, said in an announcement. “Nvidia helps these pioneers push the boundaries of quantum computing research.”
PSNC's QPUs will enable researchers to explore biology, chemistry and machine learning using two PT-1 quantum photonics systems. The systems use individual photons or packets of sunshine at telecommunications frequencies as qubits. This enables a distributed, scalable and modular quantum architecture using off-the-shelf telecommunications components.
“Our collaboration with ORCA and Nvidia has allowed us to create a singular environment and construct a brand new quantum-classical hybrid system at PSNC,” said Krzysztof Kurowski, CTO and deputy director of PSNC, in an announcement. “Open, easy integration and programming of multiple QPUs and GPUs, efficiently managed through user-centric services, is critical for developers and users. This close collaboration paves the best way for a brand new generation of quantum accelerated supercomputers for a lot of revolutionary application areas, not tomorrow, but today.”
The QPU integrated into Jupiter will enable JSC researchers to develop quantum applications for chemical simulations and optimization problems and exhibit how classical supercomputers could be accelerated by quantum computing. It consists of superconducting qubits or electronic resonance circuits.
which could be made to behave like artificial atoms at low temperatures.
“Quantum computing is brought closer by hybrid quantum classical accelerated supercomputing,” Kristel Michielsen, head of the quantum information processing group at JSC, said in an announcement. “Through our ongoing collaboration with Nvidia, JSC researchers will advance the fields of quantum computing and chemistry and materials science.”
CUDA-Q is an open-source and QPU-agnostic quantum classical accelerated supercomputing platform. It is utilized by most corporations deploying QPUs and offers top-notch performance.
Nvidia's Grace Hopper superchip attacks climate change
Regarding the Nvidia Grace Hopper superchips within the nine supercomputing centers, Nvidia said this move will speed up scientific research and discovery.
New Grace Hopper-based supercomputers coming online include EXA1-HE in France from CEA and Eviden; Helios on the Cyfronet Academic Computer Center in Poland and Alps on the Swiss National Supercomputing Center of Hewlett-Packard Enterprise (HPE); Jupiter on the Jülich Supercomputing Center in Germany; DeltaAI on the National Center for Supercomputing Applications on the University of Illinois Urbana-Champaign; and Miyabi on the Japanese Joint Center for Advanced High Performance Computing – founded between the Center for Computational Sciences on the University of Tsukuba and the Information Technology Center on the University of Tokyo.
CEA, the French Commission for Alternative Energies and Atomic Energy, and Eviden, an Atos Group company, announced in April the delivery of the EXA1-HE supercomputer, based on Eviden's BullSequana XH3000 technology. The BullSequana XH3000 architecture features recent, patented hot water cooling
system, while the EXA1-HE is supplied with 477 computing nodes based on Grace Hopper.
“AI is accelerating climate change research, accelerating drug discovery, and driving breakthroughs in dozens of other areas,” Ian Buck, vice chairman of hyperscale and HPC at Nvidia, said in an announcement. “Nvidia Grace Hopper-based systems have gotten a necessary a part of HPC as they’ll transform industries while increasing energy efficiency.”
In addition, Isambard-AI and Isambard 3 from the University of Bristol within the United Kingdom, in addition to systems from Los Alamos National Laboratory and the Texas Advanced Computing Center within the United States, join a growing wave of Nvidia Arm-based supercomputers using Grace CPU superchips and use the Grace Hopper platform.
Sovereign AI
The push to construct recent, more efficient AI-based supercomputers is accelerating as countries world wide recognize the strategic and cultural importance of sovereign AI and put money into domestic and hosted data, infrastructure and workforce to fuel innovation.
Combining Arm-based Nvidia Grace CPU and Hopper GPU architectures using Nvidia NVLink-C2C interconnect technology, GH200 serves because the engine behind scientific supercomputing centers world wide. Many centers plan to maneuver from system installation to real science inside a couple of months
as a substitute of years.
Isambard-AI phase one consists of an HPE Cray Supercomputing EX2500 with 168 Nvidia GH200 superchips, making it one of the efficient supercomputers ever built. When the remaining 5,280 Nvidia Grace Hopper superchips arrive on the University of Bristol's National Composites Center this summer, performance will probably be increased by around 32 times.
“Isambard-AI positions the UK as a worldwide leader in AI and can help drive open science innovation at home and abroad,” Simon McIntosh-Smith, a professor on the University of Bristol, said in an announcement. “Working with Nvidia, we accomplished the primary phase of the project in record time, and when it’s accomplished this summer, there will probably be an enormous leap in performance to advance data evaluation, drug development, climate research and plenty of other areas.”
