The quantum computing company Quantinuum recently announced a quantum computer it says outperformed a landmark Google computer’s result 100-fold.
The 2019 Google result used a specific test called the linear cross entropy benchmark in the attempt to demonstrate quantum supremacy, the point at which quantum computers outperform state-of-the-art ordinary (or classical) computers.
What exactly is a quantum computer?
Quantum computers operate on quantum bits. Quantum bits (qubits for short) are like ordinary computer bits, except their values can be both 0 and 1 simultaneously. Thanks to this quantum quirk, the computers can consider more solutions to a problem faster than a classical computer. Eventually, quantum computers should be able to solve problems that classical computers cannot.
But quantum computers don’t look like ordinary computers. That’s because their qubits are often supercooled atoms, set up in an array. Cooled to such a degree, the atoms enter a quantum state. The moment any one of the qubits’ value is certain, the quantum state decoheres and the quantum operation falls apart. For that reason, quantum computers as they currently exist are only in dedicated research and laboratory settings.
What did the Quantinuum quantum computer do?
The Quantinuum computer outperformed a significant 2019 achievement by Google’s Sycamore processor, which took about 200 seconds to perform a task that would take a state-of-the-art classical supercomputer about 10,000 years.
To achieve the result, the Quantinuum team upgraded its H2-1 processor from a 32-qubit system to a 56-qubit system, vastly increasing its computing power. According to a Quantinuum release, its quantum computer also ran its algorithm with about 30,000 times less power than it would’ve taken a classical computer to run the operation.
Importantly, the Quantinuum computer achieved a new record for the cross entropy benchmark, a metric used to compare the performance of different quantum computers. The benchmark measures the power of the quantum system; the noisier the system, the worse (closer to zero than 1) your results are. Google’s 2019 score on the cross entropy benchmark was ~.002; H2-1’s score was ~.35. “In contrast to past announcements associated with XEB experiments, 35% is a significant step towards the idealized 100% fidelity limit in which the computational advantage of quantum computers is clearly in sight,” stated a Quantinuum release. The team’s research is currently hosted on preprint server arXiv.
What else do quantum computers do?
Quantum computers are testbeds for the future of information—that is, the way that people store and move data, as well as compute new information. Last year, a different team of researchers showed how quantum computers could run computations in a way that looks a lot like time travel.
“The experiment that we describe seems impossible to solve with standard (not quantum) physics, which obeys the normal arrow of time,” David Arvidsson-Shukur, a quantum physicist at the University of Cambridge and the study’s lead author, told Gizmodo at the time. “Thus, it appears as if quantum entanglement can generate instances which effectively look like time travel.”
The previous year, another team claimed they managed to create a quantum wormhole—a portal through which quantum information could instantaneously travel.
Quantinuum has also run the news circuit (no pun intended). In 2022, a team using a Quantinuum computer managed to create a new phase of matter by blasting the qubits with lasers reading out the Fibonacci sequence.
Quantum computing sometimes reads like science fiction, because it seems so odd take advantage of the realm beyond classical physics to make complex calculations. But the systems keep getting better, and their applications are diverse (though some verge on pipe dreams). For now it’s relegated to research settings, but quantum computers are slowly creating the world of tomorrow today.