Blueprint for large-scale quantum computing with biased noise qubits and the XZZX surface code

Large-scaling quantum computers will require error correction to counteract the effect of noise.

For realistic amounts of noise, the hardware overhead for surface code error correction is

dauntingly large, with each logical qubit potentially requiring thousands of physical. One

promising approach is to use biased-noise qubits, such as Kerr cats, in which the probability of X

or Y errors is suppressed compared to Z errors. Using biased-noise qubits means we can more

effectively correct errors and allows for modifications of the surface code which reduce the

hardware overhead. The commonly considered modifications are the XY surface code, the XZZX

surface code, and the thin surface code.

In this talk, I’ll present a blueprint for building a large-scale quantum computer using the XZZX

surface. I’ll explain how to efficiently lay out the surface code, perform lattice surgery, and

distill magic states in the presence of dephasing noise. Some of these optimizations can also be

realized in the XY and thin surface code, while some are specific to the XZZX surface code.

Finally, I will derive estimates for the improved overhead offered by the XZZX surface code, and

compare the overhead to the XY and thin surface codes.