Fig. 3: Performance comparison of the traditional method with our approach in handling defects.
From: Low-overhead defect-adaptive surface code with bandage-like super-stabilizers
a An example with a code size of L = 7. Three highlighted circles represent potential defects, labeled as A, B, and C. We examine three scenarios: only defect A, defects A and B, and defects A, B, and C. b Comparison of logical error rate between bandage-like (B) and traditional (T) methods for the three scenarios in (a) at a physical error rate of p = 0.002. The bandage-like method shows significant advantages for both \({\left\vert 0\right\rangle }_{L}\) and \({\left\vert +\right\rangle }_{L}\) states. c–f display statistics for the bandage-like and traditional methods regarding (c) Average X distance, (d) Average Z distance, (e) Average disabled qubit percentage, and (f) Average super-stabilizer weight across different defect rates. Each data point is based on 100 generated devices with defects randomly distributed. In our simulations, defect rates are uniform for qubits and couplers. The bandage-like method consistently demonstrates substantial advantages, regardless of defect rates or code size. These advantages increase significantly with larger code sizes and defect rates.
