arXiv:2609.18922
Exact logical error rates for magic state cultivation
Kwok Ho Wan, Ainhoa Zapirain
Abstract
We compute exactly the acceptance and logical error rates for the distance $d=3$ and $d=5$ magic state cultivation circuits from Clifft [arXiv:2604.27058] and SOFT [arXiv:2512.23037] using Pauli propagation and binary tensor contraction. Actual $T$-gates are studied, not the $S$-gate proxy used for sampling. The calculation includes every fault order at several circuit-level noise strengths ($p$). We provide a series expansion form to the logical error rates, through order $(p/(1-p))^{10}$. The analytical results recover the numerical values from Clifft and SOFT at both $d=3$ and $d=5$ to within their sampling uncertainty. Then, we show that the $d=3$ and $d=5$ circuits actually have fault distances of $d_{\text{fault}}=2$ and $d_{\text{fault}}=3$ respectively, explaining the similar distance degrading effects from the companion code of [Quantum 10, 2134 (2026)].
AI Summary
Magic-state cultivation error estimates are difficult to sample at logical error rates near $10^{-9}$, and replacing $T$ gates with an $S$-gate proxy may change which faults survive post-selection. This paper computes acceptance and conditional logical error probabilities for specified $d=3$ and $d=5$ cultivation circuits with actual $T$ gates, under uniform circuit-level depolarising noise with measurement and reset flips. The calculation includes every fault order at selected noise strengths and produces exact rational series coefficients through $x^{10}$, where $x=p/(1-p)$. At $p=10^{-3}$, the logical error rates are $8.982790954\ldots\times10^{-7}$ for $d=3$ and $3.328425371\ldots\times10^{-9}$ for $d=5$; the latter has acceptance probability $0.143955584096\ldots$. The leading terms are $(32/75)x^2$ and $(574/375)x^3$, establishing circuit fault distances of two and three rather than the underlying code distances of three and five. Explicit accepted two- and three-fault patterns demonstrate these failures. Published actual-$T$ simulation intervals agree with the exact results, while the compared $d=5$ $S$-gate proxy underestimates the error rate by approximately $9.6$ at $p=10^{-3}$. Pauli propagation and stabiliser-equivalent error grouping compile non-Clifford checking stages into response tables. Binary tensor contraction joins these to injection and noisy code growth; modular arithmetic reconstructs exact probabilities, while polynomial arithmetic yields the series. The construction exploits circuit-specific structure, including syndrome dephasing during growth, and depolarising covariance. Results use the specified unitary-injection circuits and noiseless terminal readout; escape, grafting and decoding stages are excluded.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- v3 - expanded section on Pauli semiwebs and implementation available at: https://github.com/kh428/exact_ler_for_msc