Simulating time evolution on distributed quantum computers

Abstract

We study a variation of the Trotter-Suzuki decomposition, in which a Hamiltonian exponential is approximated by an ordered product of two-qubit operator exponentials such that the Trotter step size is enhanced for a small number of terms. Such decomposition directly reflects hardware constraints of distributed quantum computers, where operations on monolithic quantum devices are fast compared to entanglement distribution across separate nodes using interconnects. We simulate non-equilibrium dynamics of transverse-field Ising and XY spin chain models and investigate the impact of locally increased Trotter step sizes that are associated with an increasingly sparse use of the quantum interconnect. We find that the overall quality of the approximation depends smoothly on the local sparsity and that the proliferation of local errors is slow. As a consequence, we show that fast local operations on monolithic devices can be leveraged to obtain an overall improved result fidelity even on distributed quantum computers where the use of interconnects is costly.