IQM and AMD Demonstrate Reproducible Hybrid QPU-GPU Chemistry Workflow Using ROCm
quantumcomputingreport.com Nov 18, 2026

IQM and AMD Demonstrate Reproducible Hybrid QPU-GPU Chemistry Workflow Using ROCm

AI-summarised brief · reviewed before publication

IQM and AMD showcased a hybrid quantum‑classical chemistry workflow that splits molecular ground‑state energy calculations between IQM’s 20‑qubit Garnet superconducting processor and AMD Instinct GPUs running ROCm. Using Quantum‑Selected Configuration Interaction, the pipeline bypasses exponential memory scaling by sampling dominant electron configurations on the quantum device, filtering results on the GPU, and performing O(N³) diagonalization with JAX on ROCm. The team validated the process on H₂ in a minimal basis, reproducing its dissociation curve across 17 bond lengths on physical hardware. The open‑source notebook is available on IQM Academy and AMD Developer Cloud.

💡 Why It Matters

  • · The collaboration demonstrates a practical, multi‑vendor approach to integrating quantum processors with mainstream GPU accelerators, offering supercomputing centers a scalable path toward hybrid quantum‑classical simulations without relying on proprietary ecosystems.