This is how Microsoft wants to solve a big problem with quantum computers

Microsoft Not only invest in artificial intelligence, but also keep an eye on the future of computing. This week the company released a file Planas it will happen in the next ten years Quantum supercomputer wants to build.

It wasn’t until mid-June that IBM made a (self-described) “breakthrough” with its quantum computer (future area reported). For the first time, researchers have successfully used a conventional computer in Classic simulator to bypass. There is always news that quantum computers are faster than conventional computers. To date, however, computing tasks have been designed specifically for quantum computers.

Quantum computers are very error prone

Microsoft indicates in its roadmap that quantum computing will move on from “noise” physical qubit Very reliable logical qubits must pass. Physical qubits are not suitable for scalable applications. Today’s quantum computers generate a large number of errors, which means that calculations “make noise” and the performance of such computers cannot be fully exploited.

Faultability lies in the nature of qubits. These are very unstable and vulnerable to interference from the environment. Therefore ‘recovering’ these errors is an important task for quantitative researchers.to calculate“.

Majorana Cubitt to help

Microsoft wants to combine hundreds to thousands of physical qubits into a logical qubit that is no longer error-prone. However, this only works if the error rates of the physical qubits are below a certain threshold threshold lying down. Once this threshold of stability is reached, it is possible to manufacture reliable logical qubits.

It was said that only last year Microsoft was able to create a qubit from a file Majorana fermion to succeed. This kind of elementary physics particle is very stable, but also very difficult to produce.

Microsoft estimates that at least the first quantum supercomputer One million reliable quantum operations per second It must be done in order to be able to make valuable contributions to the solution of scientific problems. The error rate should be no more than 1 error per trillion operations. For more complex simulations in chemistry or materials research already Billions of operations per second necessary.

But we are still far from that day. Microsoft wants to reach this milestone not in decades, but in a few years.