Honeywell has issued a press release announcing they have reached a Quantum Volume metric of 64 in the latest iteration of their six qubit ion trap design. Based upon reaching this milestone, they proclaimed that they had the world’s highest performing quantum computer. Quantum Volume is a measurement created by IBM to capture the capabilities of a quantum computer based upon the number of qubits, the error rates in the qubits, the connectivity of the qubits and other factors.
However, we recommend to our readers that they take this announcement with a note of caution. The claim of having the world’s highest performing quantum computer is based solely on the fact that IBM previously announced last year that they had achieved a Quantum Volume measurement of 32. No other vendor has announced a Quantum Volume measurement for their machines. We track 87 different organizations working on quantum computing hardware and Quantum Volume information is not available from the other 85. Part of the reason may be that some companies do not believe that Quantum Volume is the best way to measure how good a quantum computer is and these companies never bothered to measure it. So how does Honeywell know for sure that theirs is the best? So, in our view, it is premature for them to state they have the world’s highest performing quantum computer.
Although the Quantum Volume test is useful because it requires a worst case configuration that forces a computer designer to equally weigh the importance of the number of qubits and the quality of those qubits. But it is an artificial measure and may not be a good determinant of how close a quantum computer is to achieving Quantum Advantage, which in our view is the real measure of a quantum computer’s success. Quantum Advantage is defined as the ability to provide a commercially valuable solution to a real world problem that cannot be reasonable achieved using classical computing technology. The best way of measuring how close a quantum computer is to reaching Quantum Advantage is through application level benchmarks. And the results will be different depending upon which application is most important to you. There are several organizations including the IEEE, the QED-C, and others looking at the best ways of benchmarking quantum computers. But it is a complex topic and a lot work still needs to be done before we can have measures that the quantum community feels comfortable with.
With all that being said, Honeywell should be congratulated for their engineering achievement. Although their latest chip has only six qubits the gate fidelities and coherence times are very good. The technology includes all-to-all coupling which can provide significant advantages and lower gate counts for many circuits. One unique innovation from Honeywell that we find interesting is the ability to measure a single qubit and branch into a different quantum program depending upon the result. This conditional execution capability could have a significant impact in the future as programmers learn how to take advantage of it. It is made possible by the long coherence times inherent in the ion trap technology and is probably not possible to do with other designs based upon superconducting technology.
The next steps will be to scale up the technology so that it can support many more qubits while increasing or perhaps maintaining the qubit quality levels. There still is a lot of development left to do and we look forward to hearing more about future developments as they work towards the goal of achieving Quantum Advantage. But it will take a lot more than 6 qubits to get it done!
June 20, 2020