A commercially-available “quantum computer” has been on the market since 2011, but it’s controversial. The D-Wave machine is nothing like other quantum computers, and until recently, scientists have doubted that it was even truly quantum at all.
But the company has released an important new result, one that in part realises Richard Feynman’s initial dreams for a quantum computer.
Scientists from D-Wave announced they have simulated a large quantum mechanical system with their 2000Q machine — essentially a cube of connected bar magnets.
The D-Wave can’t take on the futuristic, mostly non-physics-related goals that many people have for quantum computers, such as finding solutions in medicine, cybersecurity and artificial intelligence. Nor does it work the same way as the rest of the competition. But it’s now delivering real physics results. It’s simulating a quantum system.
In regular computers, any problem gets translated into arrays of controllable, two-choice switches, or bits. In quantum computers, instead, problems are translated into qubits. These two-choice switches obey the rules quantum mechanics, so that single switches can take on a combination of states and two or more distant switches can become entangled such that their joint outcomes are reliant on one another. The final answer is still just zeroes and ones.
D-Wave’s computers don’t quite work that way though. Instead, it’s as if all of these quantum switches are tiny bar magnets with a programmable electric field and magnetic current to control them. Solving problems on these computers involves giving the qubits an initial value, watching them evolve in the field over time, and then seeing what comes out.
In the past, some have criticised D-Wave for claiming that its computer could do more than it really could. Now D-Wave has shown that its machine really is quantum, but limited in its functionality.
The D-Wave computer isn’t a universal quantum computer. In the future, such a computer could break the codes that encrypt public data, and would be fully programmable. The D-Wave, meanwhile, is a very noisy system that can’t run algorithms in the way the competition’s computers can.
But in this case, it didn’t matter. D-Wave scientists found a physics problem that their machine was well-suited to study: Magnetic materials, unsurprisingly.
Here, the scientists simulated a large quantum system — meaning that even though their computer’s qubits are loops of superconducting wire, they could use it to predict the behaviour of an entirely different system, a cube of particles in a magnetic field.
“These results are a validation that the processors that we have implemented contain the physics that we said was there all along,” D-Wave scientist Richard Harris told Gizmodo.
“What we have shown is that one can embed a very different quantum system into our processor and it still works. Thus, D-Wave processors are both quantum mechanical and flexible enough to simulate other systems.”
You might be familiar with the property of “spin”, which turns particles into the smallest bar magnets. You could imagine a cubic crystal made from particles with these two-choice spins. You could then imagine these spins taking on one of three magnetic phases: Random, aligned with an external magnetic field, or alternating.
The scientists mapped the D-Wave’s qubits onto a cubic crystal eight particles to a side, and simulated how it would switch between these magnetic phases. They published their result in the journal Science.
This might sound esoteric, but it’s interesting to physicists and folks in the quantum computing field because it’s really simulating quantum mechanics working on a large scale, as devised by Richard Feynman in the early 1980s.
“This sort of simulation is something that their architecture is well-suited to do,” Aram Harrow, MIT physics professor not involved in the study, told Gizmodo.
“The authors deserve credit for making this creative connection. It is important also because it’s a much larger quantum simulation than have been run on smaller universal quantum computers.”
“I think it is the most exciting result that D-Wave has produced to date and it basically fulfils Feynman’s dream of a quantum simulator,” another physicist, Helmut Katzgraber from Texas A&M, told Gizmodo.
But Katzgraber, Harrow and even Harris all told me the same thing: That the D-Wave is not a general quantum simulator, and is best at simulating systems such as these — collections of particles with two quantum states in a magnetic field.
Ultimately, simulating quantum physics is one of several goals of quantum computers. Others hope quantum computers’ unique architectures might be useful for artificial intelligence, cybersecurity, modelling how complex molecules behave to create new medicines, and other difficult problems.
D-Wave computers are for sale, sure, but they aren't being used for any of these purposes just yet. Mainly, folks are using them to research non-physics problems that might be solved better on a D-Wave’s quantum computer than a regular computer.
But simulating a complex physics system is an important start — and Feynman would be pleased.