The D-Wave quantum computer successfully models a quantum system

D-Wave material has always occupied a unique space in the computer landscape. It is a general purpose computer that relies on quantum mechanical effects to perform calculations. And, while other quantum computer manufacturers have struggled to collect more than a few dozen qubits, D-Wave systems have already reached more than 2,000 addressable bits. But D-Wave systems do not perform the calculations in the same way and, despite all these bits, have not clearly demonstrated performance that can outperform even traditional computer hardware.

But D-Wave released a research paper in Science which suggests that the system can do some interesting things even in its current state. The company researchers decided to model a quantum system that closely resembles the bits used in the hardware itself, allowing them to examine quantum phase transitions. Although it is not yet a peak performance, it allows researchers to fully control the physical parameters of a relevant quantum system by undergoing phase changes.

Spins and spin glass

D-Wave systems can be considered as a large collection of magnets, each of which can reverse the orientations. These are not qubits in the same way as the components of the IBM or Intel quantum processors, but they depend on quantum behavior to perform calculations. On their own, there is nothing that favors one orientation over another. But put a second magnet nearby and the two influence each other; now, if we return its orientation, it changes the energy content of the system. The current D-Wave system scales this up to 2048 individual magnets, with the associated control hardware that determines which of these magnets is connected and what is the strength of that connection.

The hardware can act as a universal computer because it is possible to code the solution to the problem as the system's minimum energy state.

In this case, the system that they model is strangely similar to the D-Wave computer itself. Called a "cross-field Ising pattern", it's a cubic arrangement of magnets that can return. If these magnets are ordered so that they alternate orientations when you move in one of three dimensions, an anti-ferromagnetic is formed. But it is also possible to have configurations in which the orientations are disordered, forming what is called a "spin glass" (the magnetic properties emerge from the spin of the particles). While spin glasses are disordered, they have well-defined energies, including a low-energy state.

While the individual magnetic bits in a D-wave system are largely in one plane, it is possible to control the connections between them so that the system accurately simulates the behavior of a trellis three-dimensional. On the current generation of systems, the largest lattice that goes into the processor is one that is a cube with eight magnets on one side

Getting an answer is not really an innovative job. "An 8 x 8 x 8 trellis is the state of the art for the 1990 computer models," said Richard Harris, a D-Wave scientist who was the lead author of the paper's research. But for the company, the work represents a significant progress. "It's the first time that respectable competition has to be a multi-core server, as opposed to something like a pencil," Harris told Ars.

But, because these problems were solved previously, they provide important validation of the D-wave system. As they produce the same responses, they show that the system can correctly configure the physical system being modeled (and extract a result). "It's clear that's what we said," Harris said. "It contains the physics necessary to represent this model." Or, as he says in the document, the D-Wave material "can be used as a material physics simulator. The ability to manipulate spins and individual links allows one to explore the parameters of the material." order. "

This is also important because we already know a lot of other issues that can be mapped to these cross-field Ising models, which means that it confirms that D-Wave hardware can potentially be used to solve a wide range of problems

Performance Issues

Of course, the problem remains that of performance. Although the number of addressable bits in D-Wave hardware has increased dramatically, it remains competitive with multi-core hardware and we routinely throw clusters in case of physics issues.

But these results are also promising. Calculations on a D-Wave computer support a process called annealing. This involves putting the computer in a simple configuration and in its energetic ground state, then gently moving it to a configuration where the answer to a question can be read. If the process is done gently enough, the system stays in the ground state all the time – this naturally corresponds to a wide variety of minimization problems.

With this class of problems, researchers can explore the annealing process in more detail. system where we know what the answer should be. This will allow them to examine how the system responds by adjusting the connections between its components and exploring what happens when the system undergoes phase transitions. "We want to see how the system moves through phase transitions." Harris told Ars. "What does it look like, how can it be optimized, if it can not [be optimized] be worked?"

And, if D-Wave better understands the annealing process, it might allow to speed up their material or detect when things went wrong, and the annealing will produce an incorrect note. Leading this home, the company's CEO, Vern Brownell, went into conversation here to say, "This is helping to design the next generation of D-Wave systems."

The challenge for D-Wave is to get its next generation done in a timely manner. The company has already extended its computers to far more bits than the IBM or Intel quantum processors and is expected to maintain its lead in the foreseeable future. But this advance is not enough to allow D-Wave to outpace traditional computers for the moment. There is every reason to believe, however, that the computers on which IBM and Intel work offer massive accelerations with much less qubits than D-Wave. So, this is the stuff of an intriguing technological race.

Science 2018. DOI: 10.1126 / science.aat2025 (About DOIs).