D-Wave’s hardware has always occupied a unique space on the computing landscape. It’s a general-purpose computer that relies on quantum mechanical effects to perform calculations. And, while other quantum-computer makers have struggled to put more than a few dozen qubits together, D-Wave’s systems have already scaled to more than 2,000 addressable bits.
But D-Wave has come out with a research paper in that suggests that the system can do interesting things even in its current state. The company’s researchers have set it loose modeling a quantum system that closely resembles the bits used in the hardware itself, allowing them to examine quantum phase transitions. While this still isn’t cutting-edge performance, it does allow researchers full control over the physical parameters of a relevant quantum system as it undergoes phase changes.
Spins and spin glass
D-Wave’s systems can be thought of as a large collection of magnets, each of which can flip orientations. These aren’t qubits in the same way that the components of IBM or Intel’s quantum processors are, but they do rely on quantum behavior for performing calculations. On their own, there’s nothing that favors one orientation over another. But put a second magnet nearby and the two influence each other; now, if one flips its orientation, it changes the energy content of the system. D-Wave’s current system scales this up to 2,048 individual magnets, along with associated control hardware that determines which of these magnets is connected and how strong that connection is.
The hardware can act as a general-purpose computer because it’s possible to encode the solution to the problem as the minimal-energy state of the system.
In this case, the system they are modeling looks suspiciously like the D-Wave computer itself. Called a “transverse-field Ising model,” it’s a cubic arrangement of magnets that can flip. If these magnets are ordered such that they alternate orientations as you move in any of the three dimensions, an anti-ferromagnet is formed. But it’s also possible to have configurations in which the orientations are disordered, forming what’s called a “spin glass” (magnetic properties emerge from the spin of particles). While spin glasses are disordered, they do have well-defined energies, including a low-energy state.
While the individual magnetic bits in a D-wave system are largely in a single plane, it’s possible to control the connections among them so that the system accurately simulates the behavior of a three-dimensional lattice. On the current generation of systems, the largest lattice that will fit in the processor is one that’s a cube with eight magnets on a side.
Getting an answer is not exactly groundbreaking work. “An 8 x 8 x 8 lattice is state of the art for 1990 computer models,” said Richard Harris, a D-Wave scientist who was lead author on the research paper. But for the company, the work represents significant progress. “It’s the first time the respectable competition has to be a multicore server, as opposed to something like a pencil,” Harris told Ars.
But, because these problems have been solved previously, they provide an important validation of the D-Wave system. Since they produce the same answers, they show that the system can properly set up the physical system being modeled (and extract a result from it). “It’s clear evidence that it is what we said it was,” Harris said. “It contains the necessary physics to represent this model.” Or, as he put it in the paper, D-Wave’s hardware “can be used as a material physics simulator. The ability to manipulate individual spins and bonds allows one to explore order parameters in ways that are not possible in bulk condensed matter systems.”
This is also significant because we already know lots of other problems that can be mapped onto these transverse-field Ising models, which means that this is a confirmation that the D-Wave hardware can potentially be used to solve a wide range of problems.
Of course, the problem remains one of performance. While the number of addressable bits in D-Wave’s hardware has gone up considerably, it’s still only competitive to multicore hardware, and we routinely throw clusters at physics problems.
But there’s some promising news for performance in these results, too. Calculations on a D-Wave computer rely on a process called annealing. This involves putting the computer in a simple configuration and in its energetic ground state, then moving it gently to a configuration where the answer to a question can be read out. If the process is done gently enough, the system stays in the ground state the whole time—this naturally maps to a wide variety of minimization problems.
With this class of problems, researchers can explore the annealing process in more detail in a system where we know what the answer should be. This will let them examine how the system responds as they tweak the connections among its components and explore what happens as the system undergoes phase transitions. “We want to look at how the system moves through the phase transitions.” Harris told Ars. “What does it look like, how can it be optimized? If it can’t [be optimized], can it be worked around?”
And, if D-Wave gets a better understanding of the annealing process, it could enable them to run their hardware faster or detect when things have gone wrong, and the annealing is going to produce an incorrect note. Driving this home, the company’s CEO, Vern Brownell, broke into the conversation here to say, “This is helping design the next generation of D-Wave systems.”
The challenge for D-Wave is to get its next few generations done in a timely manner. The company has already scaled its computers up to far more bits than IBM’s or Intel’s quantum processors, and it should retain its lead for the foreseeable future. But that lead isn’t enough to allow D-Wave to outpace traditional computers at the moment. All indications are, however, that the sorts of computers IBM and Intel are working on will provide massive speedups with far fewer qubits than D-Wave already has. So, this has the makings of an intriguing technology race.