Edit on GitHub

Qecsim.jl - Quantum Error Correction Simulator

Introduction

Qecsim.jl is a Julia package for simulating quantum error correction using stabilizer codes.

NOTE: Qecsim.jl is a ground-up rewrite of the Python package qecsim.

Installation

Qecsim.jl is installed, like any other registered Julia package, using the Julia package manager Pkg:

pkg> add Qecsim  # Press ']' to enter the Pkg REPL mode.

or

julia> using Pkg; Pkg.add("Qecsim")

Examples

Simulation run

julia> using Qecsim, Qecsim.BasicModels, Qecsim.GenericModels
julia> data = qec_run(FiveQubitCode(), BitFlipErrorModel(), NaiveDecoder(), 0.1; max_runs=100);┌ Info: qec_run: starting
│   code = Qecsim.BasicModels.BasicCode(["XZZXI", "IXZZX", "XIXZZ", "ZXIXZ"], ["XXXXX"], ["ZZZZZ"], (5, 1, 3), "5-qubit")
│   error_model = Qecsim.GenericModels.BitFlipErrorModel()
│   decoder = Qecsim.GenericModels.NaiveDecoder(10)
│   p = 0.1
│   random_seed = nothing
│   max_runs = 100
└   max_failures = nothing
[ Info: qec_run: rng=MersenneTwister(0x84c2b3c172c3851d38504b11f4314658)
[ Info: qec_run: complete: data=Dict{Symbol, Any}(:error_weight_pvar => 0.4299, :time_steps => 1, :n_logical_commutations => [7, 3], :error_weight_total => 51, :wall_time => 0.001950939, :n_k_d => (5, 1, 3), :error_model => "Bit-flip", :physical_error_rate => 0.102, :measurement_error_probability => 0.0, :error_probability => 0.1, :n_success => 93, :logical_failure_rate => 0.07, :custom_totals => nothing, :code => "5-qubit", :decoder => "Naive", :n_fail => 7, :n_run => 100)
julia> dataDict{Symbol, Any} with 17 entries:
  :error_weight_pvar             => 0.4299
  :time_steps                    => 1
  :n_logical_commutations        => [7, 3]
  :error_weight_total            => 51
  :wall_time                     => 0.00195094
  :n_k_d                         => (5, 1, 3)
  :error_model                   => "Bit-flip"
  :physical_error_rate           => 0.102
  :measurement_error_probability => 0.0
  :error_probability             => 0.1
  :n_success                     => 93
  :logical_failure_rate          => 0.07
  :custom_totals                 => nothing
  :code                          => "5-qubit"
  :decoder                       => "Naive"
  :n_fail                        => 7
  :n_run                         => 100

Simulation plot

NOTE: This example assumes that Plots is installed.

julia> using Qecsim, Qecsim.BasicModels, Qecsim.GenericModels, Logging, Plots
julia> with_logger(NullLogger()) do
           error_probabilities = 0.0:0.02:0.5
           codes = [FiveQubitCode(), SteaneCode()]
           error_model = DepolarizingErrorModel()
           decoder = NaiveDecoder()
           max_runs = 5000
           failure_rates = [[] for _ in codes]
           for p in error_probabilities, (code, f) in zip(codes, failure_rates)
               data = qec_run(code, error_model, decoder, p; max_runs=max_runs)
               push!(f, data[:logical_failure_rate])
           end
           labels = reshape([label(c) for c in codes], 1, :)
           plot(
               error_probabilities, failure_rates; label=labels, legend=:right,
               xlabel="Error probability", ylabel="Logical failure rate"
           )
       end;

examples plot

Citing

Please cite Qecsim.jl if you use it in your research. It was first introduced in the following thesis:

A suitable BibTeX entry is:

@phdthesis{qecsim,
    author = {Tuckett, David Kingsley},
    title = {Tailoring surface codes: Improvements in quantum error correction with biased noise},
    school = {University of Sydney},
    doi = {10.25910/x8xw-9077},
    year = {2020},
    note = {(qecsim: \url{https://github.com/qecsim/Qecsim.jl})}
}

License

Qecsim.jl is released under the BSD 3-Clause license, see LICENSE.