Today, at the annual meeting of the American Institute of Physics, Google Quantum AI Labs announced its latest quantum processor Bristlecone, with 72 qubits, and Google said that Bristlecone's goal is to provide researchers with a test platform for 'researching quantum bits The systematic error rate and scalability of the technology, and its applications in quantum simulation, optimization and machine learning. '
An important issue that confronts all quantum computers is the error rate.Quantum computers generally operate at very low temperatures (millikelvin) and need to be guaranteed that they are unaffected by the environment because today's qubits are still not highly Stable, even a little noise will cause the system to go wrong.
72 quantum bit chip
Because of this, the qubits in modern quantum processors (that is, the 'quantum-computed version' of traditional bits) are not really single qubits, but are a combination of many traditional bits, which helps explain some potential errors. Another limiting factor is that most systems can only maintain state in 100 microseconds.
Google demo system results show that the reading error rate is 1%, the single qubit is 0.1%, and the double qubit gate is 0.6%. The industry generally believes that quantum quanta hegemony requires at least 49 qubits, but Google said that the qubit is not Quantum Computers all. 'The equipment that runs like Bristlecone with low system error requires coordination between software and control electronics to a whole set of technologies for the processor itself,' the team wrote in a blog, ' To do this, you need to carefully perform system engineering in multiple iterations.
According to a blog post by Google's Quantum AI Lab team, the team's strategy for building quantum computers is to explore recent applications using systems compatible with large-scale universal error-correcting quantum computers. In order for quantum processors to run algorithms that go beyond the classical analog range , It not only requires a large number of qubits, but it is also crucial that the processor must also have low error rates for both readout and logic operations, such as single-bit gates and double-bit gates.
Google's announcement today will put new pressure on other teams that are working on building functional quantum computers.