Classical processing for topological quantum error correction
Year: 2017
Type: dissertation
Abstract: This thesis deals with a series of issues implementing the classical software stack to control a large-scale quantum computer. Due to the relatively high error rates of quantum hardware expected for the foreseeable future, high threshold error correction schemes such as the topological cluster state and the surface code, will be essential to performing robust quantum computation. The work begins by bringing new functionality to the Autotune software library to support the presence of runtime qubit loss when using the topological cluster state for error correction. Through simulation, it is shown that <1% runtime loss is required for minimal space-time overhead. The thesis continues by constructing new software, LogiQSim, which enables fully-fault-tolerant quantum simulation and error analysis of logical quantum gates. Through the implementation of byproduct operator tracking and the placement of defects and correlation fragments, it was possible to demonstrate a full quantum simulation of a working logical controlled NOT gate (CNOT), as well as estimate its logical error rate directly through simulation. This approach was further improved to allow the simulation of the Clifford portion of A-state distillation, significant as this circuit is the predominant source of overhead in quantum computation when using the topological cluster state. The simulated overhead of the distillation process was compared to estimates calculated from the results of the logical CNOT simulations in order to determine the accuracy of estimating a larger circuit using the simulations of a smaller one. Furthermore, we found that the relationship between the structure of a correlation fragment and its contribution to the logical error rate is complex. This is due to an asymmetry in the frequency of logical errors that are the result of error chains connecting two defects as compared to rings of errors around defects. The work presented in this thesis makes substantial progress towards the understanding and development of a practical classical software stack that can be used to run a topologically error corrected quantum computer, and utilizes this (in the absence of physical hardware) to simulate and provide more accurate overhead estimates for key logical operations.
Author: Adam C. Whiteside
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: Australian Government
OpenAlex ID: W2793754553
