Proc. Natl Acad. Sci. USA https://doi.org/ctvz (2018)

We keep hearing about the tough problems that quantum computers will allow us to crack. As tantalizing — and potentially useful — as they sound, our experimental capabilities still relegate them to a distant future. Many have thus turned towards the more immediate goal of finding tasks beyond the reach of classical machines, but the utility of these tasks is rarely a priority. Andrew Childs and colleagues have now taken a more practical look at what lies between these extremes.

Through circuit optimization, Childs et al. characterized truly useful computations whose hardware requirements are not too far out of reach. They specifically considered the full quantum dynamics of 1D spin chains — a notoriously difficult system to simulate for any more than 20 or so spins, which is often too few to fully reveal phase transitions. The authors showed that digital quantum simulation of a 50-spin chain could be achieved by current algorithms with fewer than 100 qubits executing around 108 gates — orders of magnitude below the number needed for factoring large numbers or simulating quantum chemistry.