Fault-Tolerant One-Bit Addition with the Smallest Interesting Colour Code

TL;DR

This paper demonstrates fault-tolerant one-qubit addition on a quantum computer using an [[8,3,2]] colour code, reducing overhead and error rates compared to unencoded implementations.

Contribution

It introduces a simplified fault-tolerant implementation of a quantum addition algorithm using a small colour code, minimizing error correction overhead.

Findings

Fault-tolerant circuit has an error rate of ~1.1e-3.

Unencoded circuit has an error rate of ~9.5e-3.

Reduced two-qubit gates and measurements to 36.

Abstract

Fault-tolerant operations based on stabilizer codes are the state of the art in suppressing error rates in quantum computations. Most such codes do not permit a straightforward implementation of non-Clifford logical operations, which are necessary to define a universal gate set. As a result, implementations of these operations must either use error-correcting codes with more complicated error correction procedures or gate teleportation and magic states, which are prepared at the logical level, increasing overhead to a degree that precludes near-term implementation. In this work, we implement a small quantum algorithm, one-qubit addition, fault-tolerantly on the Quantinuum H1-1 quantum computer, using the [[8,3,2]] colour code. By removing unnecessary error-correction circuits and using low-overhead techniques for fault-tolerant preparation and measurement, we reduce the number of error-prone two-qubit gates and measurements to 36. We observe arithmetic errors with a rate of $\sim 1.1 \times 10^{-3}$ for the fault-tolerant circuit and $\sim 9.5 \times 10^{-3}$ for the unencoded circuit.