BraketSimulator

This package is a quantum circuit simulator written in the Julia programming language, meant to be compatible with the Amazon Braket SDK. It can simulate gate-based quantum circuits using both statevectors and density matrices (when noise is present).

You do not need a Python installation or the Python Amazon Braket SDK installed to use this package.

Supported operating systems and architectures:

All necessary Julia packages will be installed for you when you run Pkg.add("BraketSimulator") or ] instantiate (if you're doing a dev install). The minimum supported Julia version is 1.9.

In order to generate a new set of precompilation instructions, you'll need to comment out the include("precompile.jl") lines in src/BraketSimulator.jl and ext/BraketSimulatorPythonExt/BraketSimulatorPythonExt.jl. Then, cd to precompile and run:

julia --project=../test snoop_compilation.jl

This will run the test suite using SnoopCompile.jl to generate precompilation instructions. Once this is completed, you can uncomment include("precompile.jl") lines in src/BraketSimulator.jl and ext/BraketSimulatorPythonExt/BraketSimulatorPythonExt.jl.

BraketSimulator.jl has an optional extension to support integration with Python (BraketSimulatorPythonExt).

To use BraketSimulatorPythonExt, you will need to install PythonCall.jl, and then load it before BraketSimulator.jl like so:

using PythonCall, BraketSimulator

PythonCall.jl will try to install the necessary Python dependencies using the CondaPkg.toml present in top-level package folder. If you already have all the necessary Python dependencies installed, you can set the environment variable JULIA_CONDAPKG_BACKEND="Null" to have CondaPkg.jl use your system Python and its installed packages.

If you want to use this package from Python, see the Python wrapper package amazon-braket-simulator-v2-python, which provides a Python interaction layer and will install Julia, this package and other Julia dependencies in a sandboxed virtual environment for you.

Launch Julia with the command line option -t auto to have Julia launch and manage threads to parallelize the simulation(s). BraketSimulator.jl will parallelize both across tasks within a batch and inside a single task's evolution.

Keep in mind that the first qubit has index 0, not index 1.

In this example, we create a Greenberger-Horne-Zeilinger (GHZ) state and simulate it:

using Braket, BraketSimulator

function ghz_circuit(qubit_count::Int)
    ghz_circ = Circuit()
    H(ghz_circ, 0)
    for target_qubit = 1:qubit_count-1
        CNot(ghz_circ, 0, target_qubit)
    end
    return ghz_circ 
end

device = LocalSimulator("braket_sv_v2")
n_qubits = 10
task_result = simulate(device, ghz_circuit(n_qubits), shots=10)

See CONTRIBUTING for more information.

This project is licensed under the Apache-2.0 License.