Qu&Co comments on this publication:
In this paper, Djidjev et al. evaluate the performance of the D-Wave 2X quantum annealer on two NP-hard graph problems: clique finding and graph partitioning. Overall, they conclude that general problems which allow to be mapped onto the D-Wave architecture are typically still too small to show a quantum speedup (although the D-wave does provide similar quality solutions as the classical solvers). For simple simulated annealing algorithms, D-Wave is considerably faster and selected instances especially designed to fit D-Wave's particular chimera architecture can be solved orders of magnitude faster than with classical techniques.