- IBM integrates quantum processors with classical supercomputers for coordinated scientific computations
- Quantum-centric supercomputing enables workloads to switch between CPUs, GPUs and QPUs
- Researchers have successfully simulated complex molecules using hybrid quantum-classical workflows
IBM has outlined a new reference architecture designed to combine quantum processors with traditional supercomputing infrastructure.
The company describes the concept as quantum-centric supercomputing, an approach intended to connect quantum processing units with GPUs and CPUs in large-scale computing environments.
The architecture is designed to work across research centers, on-premises infrastructure and cloud systems, while supporting coordinated workflows between different types of hardware.
The article continues below
Design of a unified quantum-classical computing environment
The proposed design integrates quantum processors with classical computer clusters, high-speed network systems and shared storage infrastructure.
IBM says this arrangement allows scientific workloads to move between different processors depending on the computational demands of the task.
Open software frameworks, including Qiskit, are intended to manage planning and coordination across the combined systems.
Jay Gambetta, director of IBM Research, said the goal is to merge quantum and classical computing resources into a unified environment capable of solving problems that traditional supercomputers struggle to simulate.
“More than four decades ago, Richard Feynman envisioned computers that could simulate quantum physics,” he said.
“The future lies in quantum-centric supercomputing, where quantum processors work together with classical high-performance computing to solve problems that were previously out of reach.”
IBM and its research partners have reported measurable scientific results using hybrid quantum-classical computing.
Teams from the University of Manchester, the University of Oxford, ETH Zurich, EPFL and the University of Regensburg confirmed the unusual electronic structure of a half-Möbius molecule.
Cleveland Clinic researchers simulated a 303-atom tryptophan cage miniprotein, while IBM, RIKEN and the University of Chicago identified the lowest energy states of engineered quantum systems that outperformed classical methods.
In a larger experiment, an IBM quantum processor exchanged data with 152,064 classical nodes of RIKEN’s Fugaku supercomputer to simulate iron-sulfur molecular clusters, critical in biology and chemistry.
Despite these demonstrations, hybrid quantum workflows remain technically complex, as researchers often need to coordinate data transfers, scheduling, and algorithm execution between separate computing systems.
IBM’s reference architecture attempts to address these challenges through coordinated software orchestration and shared infrastructure designed to connect quantum and classical resources.
The company describes a staged development path where quantum processors will first act as specialized accelerators in existing supercomputing centers.
Later stages would involve closer coupling between quantum hardware and classical computer clusters through advanced middleware systems.
These experiments show that hybrid quantum systems can contribute to specialized scientific computations – however, the results remain largely limited to controlled research environments and very specific simulations.
The roadmap indicates progress in workflow integration and algorithm development, although practical deployment outside research institutions still seems limited for now.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds. Be sure to click the Follow button!
And of course you can too follow TechRadar on TikTok for news, reviews, video unboxings, and get regular updates from us on WhatsApp also.
