Quantum-Classical Computing via Tensor Networks

Circuit knitting offers a promising path to the scalable execution of large quantum circuits by breaking them into smaller sub-circuits whose output is recombined through classical postprocessing. However, current techniques face excessive overhead due to a naive postprocessing method that neglects potential optimizations in the circuit structure. To overcome this, we introduce qTPU, a framework for scalable hybrid quantum-classical processing using tensor networks. By leveraging our hybrid quantum circuit contraction method, we represent circuit execution as the contraction of a hybrid tensor network (h-TN). The qTPU compiler automates efficient h-TN generation, optimizing the balance between estimated error and postprocessing overhead, while the qTPU runtime supports large-scale h-TN contraction using quantum and classical accelerators. Our evaluation shows orders-of-magnitude reductions in postprocessing overhead, a $10^4\times$ speedup in postprocessing, and a 20.7$\times$ reduction in overall runtime compared to the state-of-the-art Qiskit-Addon-Cutting (QAC).