PennyLane

The iqm.qdmi.pennylane.IQMDevice adapter uses MQT Core for compilation, submission, and measurement results.

Install

uv pip install "iqm-qdmi[pennylane]"

Configure a device

Select a device from the installed catalogue:

import pennylane as qp

device = qp.device("iqm.qdmi", device_id="iqm.emerald.mock", wires=2)

Authentication defaults to IQM_TOKEN or IQM_TOKENS_FILE. Pass token, tokens_file, or calibration_set_id to configure the session explicitly. iqm.default also reads the endpoint and quantum computer environment variables listed in the Qiskit guide; named devices use their manifest settings.

Use job_parameters={"custom1": json.dumps(options)} for IQM run-request fields. The native device validates these fields and rejects shot-discarding heralding; see IQM run-request options.

Run a QNode

@qp.qnode(device)
def bell():
    qp.Hadamard(0)
    qp.CNOT([0, 1])
    return qp.sample(wires=[0, 1])


samples = qp.set_shots(bell, shots=1000)()

Supported scope

MQT Core compiles circuits for the selected device, including routing through computational resonators. Execution uses finite shots. See the MQT Core PennyLane guide for supported operations, measurements, and optional Catalyst integration.