IBM claims huge strides in quantum computing
With its latest research, Big Blue says it's reached device performance close to the minimum requirements for implementing a "practical quantum computer." But many hurdles remain.
Scientists at IBM say they have made a quantum computing breakthrough that demonstrates that a full-scaleis not only possible but is within reasonable reach.
In an announcement being made today at the American Physical Society in Boston, Matthias Steffen, manager of IBM's experimental quantum computing group, will unveil the research that has led his team to conclude they are the brink of developing scalable technology that could far outstrip what even the strongest supercomputers can do today.
A traditional bit has only two states--zero and one. But for its quantum computing efforts, IBM has decided to pursue what it's calling superconducting qubits, in large part because that physical system is compatible with microfabrication components for silicon technology, Steffen said.
Qubits could allow a quantum computer to perform millions of computations at a time, IBM said, and a single 250-qubit state would have more information bits than there are particles in the universe.
Other potential quantum computing physical systems that IBM decided not to pursue include ion traps and quantum dots, Steffen said. And he said that while no time frame is clear yet, he thinks that there will be a working useful quantum computing system in his lifetime.
Because qubits can be in both the zero state and the one state at the same time, it opens up a wide variety of possibilities, he explained, and will likely be a key for doing highly sensitive and very fast encryption and decryption work.
In its announcement, IBM said it has broken records for the error rate in elementary computations while maintaining the integrity of quantum mechanical properties in qubits. "These are very delicate systems," Steffen said. "The one state is usually encoded in the excited state of a quantum system...that can decay like an atom can emit a photon. That's an error when it happens to our qubit. The frequency that errors occur limit your device's performance. If errors are few and far between, you can implement error correction."
Now, Steffen said, IBM has found a way to reduce the error rate to the point where researchers can implement error correction schemes, a step that means as much as a ten-fold increase in performance is possible.
And because its error rate is approaching a significant low, Steffen said that IBM is close to being in a position where it can take five or 10 qubits together and begin to perform elementary operations.
Still, before IBM can hope to build an actual quantum computing system, it has to take small steps forward, Steffen suggested, like figuring out how to place five or 10 qubits on a chip, how to interface with them, and what kind of software to write for them.