D-Wave has become the fourth organization to offer free cloud access to a quantum computer behind the IBM Q Network, the NTT/University of Tokyo’s Quantum Neural Network, and the CAS-Alibaba Superconducting Quantum Computer. Others expected to offer cloud access soon, but not necessarily free and open, include Rigetti with their Quantum Cloud Services and Google with their Cirq platform. In addition, Microsoft is heavily promoting their Quantum Development Kit (QDK) software and simulation platform, but the associated quantum hardware is expected to be further out.
The Leap QAE provides not only open access to their D-Wave 2000Q, but it also includes an open source development tools suite called Ocean, interactive demos using Jupyter notebooks, documentation, educational resources, and an on-line community for collaboration and support. Their software is now open sourced and they have placed it on GitHub. They are also encouraging users to contribute to the developments and are providing users with an incentive to open source their software. The basic allocation of free time, is a one-time grant of one minute of computer run time (approximate equal to between 400 to 4000 program executions), but those users who agree to open source their software and place it on GitHub will be granted one minute of computer run time per month. Commercial customers who may have heavy usage can purchase additional computer run time at a rate of $2000 per hour.
One purpose of these various systems is to not only start training students, classical computer programmers and potential customers on how to use quantum computing, but to also get them thinking about what intractable classical computing problems might become solvable on a quantum computer. And a second goal is to start convincing these potential customers to use their quantum architecture rather one of the competing ones. Sort of like convincing someone to use an Android based cell phone instead if an iOS based iPhone.
The D-Wave architecture is much different from the gate levels approaches being taken by IBM, Google, Rigetti and others. The differences are as great as those between analog and digital classical computers. The D-Wave approach is based upon finding the minimum of an energy landscape expressed as a mathematical equation called a QUBO (Quadratic Unconstrained Binary Optimization). This can be very important for many real world applications that can formulated as optimization problems.
D-Wave does have some advantages over some of the others. First, they have more raw qubits, about 2040, in their D-Wave 2000Q machine versus the 50 or 72 in some of the others. As a result, D-Wave is arguing that this allows users to develop commercially useful applications on their machines now versus only the experimental demos that can be performed on the smaller gate level machines. On the other hand, some will suggest that other quantum computers have a higher level of qubit quality, as measured by such parameters as coherence time and that they believe that a lower number of high quality qubits will beat out a larger number of poorer quality qubits in providing better answers. D-Wave continues to work on improving both qubit quantity and quality with a 5000+ qubit machine expected sometime in 2019.
A second advantage is that it may be easier for classical computer scientists and operations researchers to intuitively understand the optimization approach used in the D-Wave machine. Many of them are very familiar with this type of problem and already use classical algorithms such as simulated annealing and other solvers to find the minimums of their functions. So mapping their logistical and other problems from an existing classical optimization solver to a D-Wave quantum annealing solution may be much easier for them than the less familiar linear algebra based approaches that would be required with a gate-level quantum computer.
A final advantage for D-Wave is that they have had machines installed and operational for a number of years. This has allowed several users to already develop over 100 example applications on the 2000Q and previous generation machines. D-Wave has developed a good following by offering access to their machines through partnerships with the Universities Space Research Association, the Creative Destruction Lab in Canada, as well as customers including NASA, Google, Lockheed-Martin, University of Southern California, Temporal Defense Systems, and various government organizations including the Los Alamos National Lab.
For more information, you can view D-Wave’s press release on their new LEAP service here and you can sign-up for access here. The D-Wave GitHub library including the Ocean SDK can be found here. In addition, for more information on D-Wave developments and other additional quantum annealing systems, you can view our previous report on the recent 2018 Adiabatic Quantum Computing Conference here.