Distinguishing Classical and Quantum Models for the D-Wave Device

Recently the question of whether the D-Wave processors exhibit large-scale quantum behavior or can be described by a classical model has attracted significant interest. In this work we address this question by studying a 503 qubit D-Wave Two device as a „black box“, i.e., by studying its input-output behavior. We examine three candidate classical models and one quantum model, and compare their predictions to experiments we have performed on the device using groups of up to 40 qubits. The candidate classical models are simulated annealing, spin dynamics, a recently proposed hybrid O(2) rotor-Monte Carlo model, and three modified versions thereof. The quantum model is an adiabatic Markovian master equation derived in the weak coupling limit of an open quantum system. Our experiments realize an evolution from a transverse field to an Ising Hamiltonian, with a final-time degenerate ground state that splits into two types of states we call „isolated“ and „clustered“. We study the population ratio of the isolated and clustered states as a function of the overall energy scale of the Ising term, and the distance between the final state and the Gibbs state, and find that these are sensitive probes that distinguish the classical models from one another and from both the experimental data and the master equation. The classical models are all found to disagree with the data, while the master equation agrees with the experiment without fine-tuning, and predicts mixed state entanglement at intermediate evolution times. This suggests that an open system quantum dynamical description of the D-Wave device is well-justified even in the presence of relevant thermal excitations and fast single-qubit decoherence.