What percentage of superconducting QPU chips are discarded after fabrication because of high error rate? Are some physical qubits significantly bad in state of the art superconducting QPUs? Are these two problems going to be a significant challenge to scaling up QPUs? How are people addressing these two problems?
Asked
Active
Viewed 39 times
2
-
I like this question! Although I suspect it’s probably pretty proprietary information. The qubits are HUGE, and I’d bet they are run on 200mm substrates with 180nm or so lithography, pretty old technology. But, for the original Sycamore processors, only 53 of the 54 qubits were working. I suspect yield per qubit is more important than yield per processor. And if enough yielding qubits are on one processor then it’s a success. – Mark Spinelli Apr 22 '23 at 01:18