Hybrid Quantum HPC Systems Boost Simulation Performance and Efficiency.

QMIO represents a novel tightly integrated hybrid High-Performance Computing and Quantum Computing system. It combines conventional computing with a Quantum Processing Unit to potentially accelerate specific computational kernels, offering improvements in speed and energy efficiency through workload offloading. The research details the system’s hardware, software and integration middleware.

The pursuit of computational advantage continues to drive innovation in high-performance computing. Researchers are now exploring the synergistic potential of combining conventional, classical systems with the emerging capabilities of quantum processors. A collaborative effort, detailed in a new publication, presents a fully integrated hybrid system designed to exploit this potential. Javier Cacheiro and Andrés Gómez from the Galicia Supercomputing Center, alongside Álvaro C Sánchez from FSAS International Quantum Center (Fujitsu) and Russell Rundle, George B Long, and Gavin Dold, alongside Jamie Friel from Oxford Quantum Circuits (OQC), describe their work in the paper, “QMIO: A tightly integrated hybrid HPCQC system”. The team details the hardware and software architecture of QMIO, alongside insights gained from its development and initial operation, offering a valuable contribution to the field of hybrid quantum-classical computation.

Hybrid Computation System Demonstrates Functional Integration of HPC and QC

Recent developments in quantum computing offer potential acceleration of complex computations beyond the capabilities of classical high-performance computing (HPC). This is driving research into hybrid architectures that effectively integrate both paradigms. This report details QMIO, a functional hybrid HPC and quantum computing (QC) system designed to tightly couple conventional HPC resources with quantum processing units (QPUs), establishing a platform for collaborative computation and exploration. The system’s architecture comprises specialized hardware, sophisticated software, and crucial integration middleware, providing insights from its design, implementation, and operational performance.

The core principle underpinning QMIO lies in strategically offloading specific computational kernels to the QPU. This leverages the potential of quantum algorithms to accelerate tasks intractable or inefficient on classical HPC systems. By selectively assigning these kernels, the system aims to enhance computational speed, improve complex simulations, and potentially reduce energy consumption. Researchers designed the system to facilitate seamless communication and data transfer between the classical and quantum domains, ensuring efficient execution of hybrid algorithms and maximising quantum acceleration benefits.

Several publications detail components and approaches contributing to the system’s functionality. Kawase et al. (2021) and subsequent work by Masumura et al. and Mitarai et al. present Qulacs, a fast and versatile quantum circuit simulator. Tools like Qulacs (Suzuki et al. 2021) and its distributed extension, mpiqulacs (Honda et al. 2023), facilitate the development and testing of quantum algorithms. Cao et al. (2023) present a novel programming model and runtime system specifically designed for hybrid HPC-QC architectures, enabling developers to express and execute hybrid algorithms easily.

Recent advancements in quantum error mitigation techniques are crucial for improving the reliability of quantum computations, addressing a key challenge in developing practical quantum algorithms. Researchers are actively exploring strategies including zero-noise extrapolation, probabilistic error cancellation, and symmetry verification, to reduce the impact of noise on quantum computations.

Integrating quantum and classical resources presents challenges related to data transfer, synchronisation, and communication. Researchers are exploring techniques including high-bandwidth interconnects, efficient data serialisation formats, and optimised communication protocols to address these challenges.

The successful development of QMIO and supporting research demonstrates the feasibility of building practical hybrid computing systems and paves the way for future advancements. Researchers are exploring more scalable and robust quantum hardware, more efficient programming models and runtime systems, and new quantum algorithms and applications.

This research highlights the importance of collaboration and interdisciplinary expertise in advancing hybrid computing, bringing together experts in quantum physics, computer science, and engineering. Continued development of hybrid computing systems requires sustained investment in research and development, alongside cultivating a skilled workforce capable of designing, building, and operating these complex systems.