bluequbit 0.18.5b1

Python SDK to BlueQubit app

BlueQubit Python SDK

BlueQubit Python SDK is a quantum computing SDK that lets you build, optimize, and run quantum circuits on BlueQubit platform from Python. Whether you need quantum simulation on large CPU/GPU devices or quantum circuit optimization at scale, this quantum development kit has you covered.

Quick Start

1. Register on https://app.bluequbit.io and copy the API token.

2. Install Python SDK from PyPI:

    pip install bluequbit

3. Import and initialize the SDK (importing also Qiskit for circuit building):

    import qiskit

    import bluequbit

    bq_client = bluequbit.init("<token>")

4. An example of how to run a Qiskit circuit using the SDK:

    qc_qiskit = qiskit.QuantumCircuit(2)
    qc_qiskit.h(0)
    qc_qiskit.x(1)

    job_result = bq_client.run(qc_qiskit, job_name="testing_1")

    state_vector = job_result.get_statevector() 
    # returns a NumPy array of [0. +0.j 0. +0.j 0.70710677+0.j 0.70710677+0.j]

5. Here is an example to run the previous code asynchronously:

    # This line is non-blocking.
    job = bq_client.run(qc_qiskit, job_name="testing_1", asynchronous=True)

    # This line blocks until the job is completed.
    job_result = bq_client.wait(job.job_id)

    # If you want to cancel a pending job.
    bq_client.cancel(job.job_id)

6. This is how you can use our mps.cpu device:

    # a 40 qubit GHZ state
    num_qubits = 40
    qc = qiskit.QuantumCircuit(num_qubits)
    qc.h(0)
    for i in range(num_qubits - 1):
        qc.cx(i, i+1)
    
    qc.measure_all()
    
    options = {
        "mps_bond_dimension": 2,
    }
    result = bq_client.run(qc, device="mps.cpu", options=options) # should take ~30seconds
    print(result.get_counts())

The mps.gpu device can be used in the same way. Note that while mps.gpu is faster, you need to have a certain amount of minimal balance to be able to use it.

7. An example of how to use our Pauli Path Simulator to compute observable expectations:

    qc = qiskit.QuantumCircuit(3)
    qc.h(0)
    qc.h(1)
    qc.h(2)
    qc.ry(0.3, 0)
    qc.rz(0.6, 1)
    qc.rx(0.7, 2)

    # observable
    pauli_sum = [
        ("XYZ", -0.5),
        ("XXX", 0.2),
        ("XII", 0.3),
        ("III", 0.4),
    ]

    options = {
        "pauli_path_truncation_threshold": 0.1,
    }
    result = bq_client.run(qc, device="pauli-path", pauli_sum=pauli_sum, options=options)
    expectation_value = result.expectation_value
    # returns 0.9411262759533625

8. An example of how to run a Pennylane circuit:

    import pennylane as qml
    from pennylane import numpy as np
    
    # use "bluequbit.cpu" device name to run on CPU
    dev = qml.device("bluequbit.cpu", wires=1, token="<token>")
    # use "bluequbit.gpu" device name to run on GPU
    # dev = qml.device("bluequbit.gpu", wires=1, token="<token>")
    
    @qml.qnode(dev)
    def circuit(angle):
        qml.RY(angle, wires=0)
        return qml.probs(wires=0)
    
    
    probabilities = circuit(np.pi / 4)
    # returns a NumPy array of [0.85355339 0.14644661]

To use the Pennylane plugin, you must have pennylane version 0.39 or above installed. bluequbit.cpu supports Pennylane circuits with up to 32 qubits (wires). bluequbit.gpu supports Pennylane circuits with up to 33 qubits (wires).

9. This SDK requires Python versions 3.10 or above. The package is tested extensively on Python 3.10.

Full reference

Please find detailed reference at https://app.bluequbit.io/sdk-docs.

Questions and Issues

Please submit questions and issues to [e-mail].