Large-Scale Quantum Chemistry Computations Using Quantum-HPC Hybrid Systems

Harnessing the potential of noisy intermediate-scale quantum (NISQ) devices requires close integration with classical high-performance computing (HPC) resources. In this work, we explore a quantum–HPC hybrid framework designed to fully exploit the capabilities of both quantum and classical resources in large-scale quantum chemistry simulations. Specifically, we utilize quantum processors as samplers that guide the classical subspace diagonalization on a supercomputer, where the classical component handles large-scale Hamiltonian construction and optimization. To enhance overall efficiency, we introduce (a) a feedback loop for circuit refinement, (b) large-scale selected subspace diagonalization leveraging massive parallelism, and (c) orbital optimization strategies. Applying this approach to iron–sulfur clusters, [2Fe–2S] and [4Fe–4S], we performed simulations involving up to 77 qubits and approximately 16,000 HPC nodes, achieving substantial improvements in the ground state energies compared to our previous studies.