IBM’s new quantum computer, Osprey, is more than three times the size of Eagle, the previous record-breaking one.
IBM has built the largest quantum computer ever, named Osprey (“osprey”), with 433 qubits, or quantum bits. That makes Osprey three times as many as the company’s previous record-breaker with 127 qubits, and even eight times more than Sycamore, Google’s 53-qubit quantum computer.
Quantum computers can solve computational problems that conventional computers cannot. But it is not yet clear what exactly is the best way to build such a machine. IBM and Google base their qubits on small rings of superconductors. Other designs use charged atoms, photons or light particles, or unusually large atoms.
READ ALSO
The silent children of the Concorde
Chip control
Building Osprey was more challenging than just making more qubits, says quantum researcher Oliver Dial from IBM. Qubits can be disrupted by the forces that neighboring qubits exert on them, similar to how accidental electrical contacts cause problems in a conventional computer. The more qubits crammed together on a chip, the greater the chance of this happening.
The qubits are controlled with conventional electronics. Dial says he and his team have designed the connections there to keep the qubits under as close control as possible.
Since the superconducting rings only work at temperatures close to -273 degrees Celsius, Osprey must be kept in a special cooler. In addition, the researchers had to ensure that the electric current in the wires does not cause an increase in temperature.
Quantum superiority
Scientists have high expectations of quantum computers. That’s because, unlike conventional bits, qubits can take on multiple values at the same time: 0, 1, and, due to the peculiarity of quantum mechanics, both 0 and 1 at the same time. This would make them smart enough to perform certain calculations that are far too complex even for today’s supercomputers. An example of this is the simulation of chemical reactions.
In 2019, Google claimed a breakthrough, when Sycamore solved a problem the company said was impossible for a conventional computer. This claim of “quantum superiority” has since been questioned repeatedly.
Osprey has a lot more qubits than Sycamore. And for quantum computers, the more qubits, the more powerful. A conventional computer would have to have more bits than there are atoms in the universe to equal the number of states Osprey can store and process information in.
Noise
‘By building Osprey, we have started exploring how big we can really make a quantum chip. We’ve learned that we can get over 400 qubits on a chip and still get the whole thing working. That was a huge lesson in engineering,” says Dial.
‘This computer shows that this superconducting technology is also scalable,’ says PeterLove, quantum researcher at Tufts University in Massachusetts, United States. ‘But this is a device with a lot of noise. And noise could well be the keyword for quantum computers in the coming years.’
Physicists speak of noise when quantum computers start to fail because the qubits lose their quantum state over time. Love says he has two big questions. First, can computers like Osprey also be made to detect and correct their own errors? And second, could IBM use Osprey to demonstrate quantum superiority?
Dial says the IBM team is well on its way to building a working quantum computer with 1121 qubits by 2023. At the same time, the team is developing a smaller chip to test a new method of ordering and connecting qubits. With this, the device would show less noise and make fewer errors. ‘We are not only working on increasing the scale, but also the quality and speed of our quantum computers,’ he says.