RIKEN and Quantinuum Demonstrate Heterogeneous Multi-QPU Workflows on Fugaku
AI-summarised brief · reviewed before publication
Researchers from Quantinuum, the University of Tsukuba, and RIKEN’s Center for Computational Science demonstrated a hybrid computing architecture that links the supercomputer Fugaku with two distinct quantum processors: Quantinuum’s Reimei trapped‑ion system and IBM’s Kobe superconducting system. Using an extended Tierkreis workflow engine, the team executed a multi‑modality quantum‑classical algorithm to compute the ground‑state energy of a 16‑site transverse‑field Ising model, achieving an error of 0.61 ± 0.36 J—substantially better than standalone VQE on Kobe (2.47 ± 0.10 J). The demonstration also validated biomolecular excited‑state workflows, automating 80 % of runtime coordination across classical chemistry packages and quantum circuit generation. The work shows that complex quantum chemistry and condensed‑matter calculations can span heterogeneous quantum hardware without direct quantum interconnects, using a classical supercomputer as the central orchestrator.
💡 Why It Matters
- · By integrating multiple quantum platforms through a unified workflow, the study removes the scalability bottleneck of fragmented glue code, enabling more efficient use of diverse quantum strengths.
- · This paves the way for practical, large‑scale quantum‑classical simulations that were previously limited to single‑architecture setups.
