D-Wave Systems hosted a seminar at SC17, the leading high performance computing mega conference. This featured something still rare in the world of quantum computing – insight from genuine end users.

Bo Ewald, President of D-Wave International, introduced the seminar with an accessible survey of the alternative approaches to creating quantum devices. This is a topic that has sometimes divided D-Wave from others in the community. His main point was one that I think most people will agree on. For the next decade the sector will be dominated by two parallel streams of endeavour. One is the tremendous stepwise engineering challenge to create a large scale, fault tolerant, programmable universal quantum computer. While in the meantime, shorter term business and scientific applications will progress based on simpler quantum simulators and, as Bo terms them, ‘partial’ gate model machines.

Denny Dahl, D-Wave Senior Research Scientist, demonstrated the basics of configuring a problem and running it remotely on a D-Wave 2000Q. Quite apart from seeing the specifics of this quantum annealing device, it was an interesting insight into the unique mind-set required to adapt a problem for an ‘adiabatic’ computational approach: the problem is mapped to a rugged landscape, as the mountain peaks of this landscape push themselves up, the system physically settles into the lowest point of the terrain; this low point is the optimisation that the user interprets as the solution. Quantum annealing is a form of adiabatic quantum computation, albeit one that makes compromises in favour of implementability.

John Sarrao, Los Alamos National Laboratory (LANL) Associate Director and perhaps D-Wave Systems’ most influential user, spoke at length. The LANL team now has an impressive track record: 22 quantum mini projects completed using their ‘Rapid Response’ framework. These span exploratory use of the D-Wave system across a wide range of potential applications: from NP-hard optimisation problems, to image recognition and terrorist network modelling. Each project is itself an innovative experience as problems are mapped to the new quantum annealing framework.

John put completely to one side the question of demonstrating quantum advantage or quantum supremacy. Instead, in the context of the current D-Wave 2X machine he operates, he poses the question “what can you do with new technology and bright staff?” He emphasises two subtle points:

• Access to such leading edge systems allows him to attract, develop and retain the right team
• Mapping practical problems to new innovative approaches spurs corresponding development in conventional implementations

Other organizations considering how they should prepare themselves for the quantum revolution should consider what they will do to compete for staff with the necessary skills. These are likely to be in short and shortening supply for the foreseeable future. Neither should the opportunity to spur thinking in conventional approaches be neglected. For many problems of great interest (not least integer factorisation and discrete logarithms) we don’t have any proof that more efficient conventional mathematical methods can’t indeed be found. It is an achievement for D-Wave’s young technology that it is already competitive in a selection of LANL test cases with established techniques in machine learning and combinatoric optimisation.

Another area where D-Wave’s experience with real users shows are the multiple interfaces it now offers to help non-specialists access its devices. These are still well short of the standard millennials have grown up to expect! But D-Wave now provides multiple different API variants, tailored to different types of problem.

Few doubt that we will soon see a quantum device of one flavour or another demonstrate limited quantum supremacy, a notable milestone in the long race to build a universal quantum computer.  Many would say that the gate model qubit technologies are currently in the pole positions, with a strong challenge perhaps expected later in the race from the topological car starting in the pit lane. While adiabatic quantum computing approaches are nearer the back of the grid in that headline race, fans who want to learn to drive now should consider what is the best way for their teams to gain experience. There is a lot of action on quantum’s rugged frontier.