en.Wedoany.com Reported - Quantum software development company Qunova Computing has signed a memorandum of understanding to officially join JHPC-quantum (Japan High-Performance Computing), a national quantum-supercomputing convergence project funded by Japan's New Energy and Industrial Technology Development Organization (NEDO). The agreement makes Qunova a formal participant in the program's test user project, approved after technical screening by the RIKEN Center for Computational Science. Among the 21 selected participating organizations worldwide, Qunova is one of only two non-Japanese entities granted direct access to Japan's flagship classical-quantum converged computing nodes.
The computing infrastructure managed by the JHPC-quantum framework is dedicated to addressing integration limitations between classical Message Passing Interface (MPI) supercomputing clusters and distributed quantum coprocessors. Traditional Variational Quantum Eigensolver (VQE) algorithms often suffer performance degradation on noisy intermediate-scale quantum (NISQ) hardware due to the large number of quantum measurements (shots) required for chemical accuracy and the overhead of variable transfer between Python development frameworks and high-performance classical nodes. To overcome these execution barriers, Qunova will deploy its proprietary Handover Iterative Variational Quantum Eigensolver (HI-VQE) algorithm.
The HI-VQE framework restructures standard hybrid processing parameters by introducing a mathematical "handover" loop that splits the computational load based on hardware efficiency. First, the algorithm uses the quantum processor as a target coprocessor to execute shallow hardware circuits, isolating the specific multi-reference electronic configurations that dominate the molecular ground state. Second, the initial high-dimensional Hamiltonian is rewritten and compressed into a simplified active space framework, reducing the accumulated quantum sampling burden. Subsequently, the transformed problem is passed back to the classical supercomputer to resolve the remaining configuration interaction, achieving an energy accuracy threshold of 1.6 mHa (chemical accuracy). The joint project will apply this hybrid stack to benchmark strongly correlated electron systems, focusing on iron-sulfur (Fe–S) clusters—a complex 40-qubit active space molecular simulation that serves as a diagnostic standard for battery design, materials informatics, and small-molecule drug discovery.
The test user program provides Qunova with allocated time on an interconnected national hardware network overseen by RIKEN and SoftBank. The computing infrastructure connects Japan's flagship classical supercomputer "Fugaku" with next-generation NVIDIA Grace-Blackwell liquid-cooled AI clusters via a high-speed, low-latency network bus. This classical layer is integrated with local quantum backends, including the on-site superconducting IBM Quantum System Two ("IBM Kobe") installed in Kobe and the high-fidelity ion trap Quantinuum platform ("Reimei") located in Wako. The JHPC-quantum project, launched in November 2023 and concluding in October 2028, is currently in the middle of its five-year research and development mission. The project uses a unified API layer to manage job queues and cross-node data exchange. By incorporating Qunova's hardware-agnostic solver into this multi-platform testbed, the initiative aims to establish a production-grade software library before the platform's scheduled commercial cloud pre-release in 2028.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
