The blog of Xeno, a slightly mad scientist
Two years ago Google and NASA went halfsies on a D-Wave quantum computer, mostly to find out whether there are actually any performance gains to be had when using quantum annealing instead of a conventional computer. Recently, Google and NASA received the latest D-Wave 2X quantum computer, which the company says has “over 1000 qubits.”
At an event yesterday at the NASA Ames Research Center, where the D-Wave computer is kept, Google and NASA announced their latest findings – and for highly specialised workloads, quantum annealing does appear to offer a truly sensational performance boost. For an optimisation problem involving 945 binary variables, the is up to 100 million times faster (108) than the same problem running on a single-core classical (conventional) computer.
Google and NASA also compared the D-Wave X2’s quantum annealing against Quantum Monte Carlo, an algorithm that emulates quantum tunnelling on a conventional computer. Again, a speed-up of up to 108 was seen in some cases.
Hartmut Neven, the head of Google’s Quantum Artificial Intelligence lab, said these results are “intriguing and very encouraging” but that there’s still “more work ahead to turn quantum enhanced optimization into a practical technology.”
As always, it’s important to note that D-Wave’s computers are not capable of universal computing: they are only useful for a small number of very specific tasks – and Google, NASA, and others are currently trying to work out what those tasks might be. D-Wave’s claim of “over 1,000 qubits” is also unclear. In the past, several physical qubits were clustered to create a single computational qubit, and D-Wave doesn’t make that distinction clear.