How to program Quandela QPUs

Reviewed on November 21, 2025

Quandela is a European leader in photonic quantum computing. Their technology relies on manipulating light (single photons) through linear optical components to perform quantum calculations.

Unlike superconducting qubits or cold atoms, Quandela's photonic approach uses "qumodes" and photons that circulate in optical circuits at room temperature (although emitters and detectors require cooling). This offers a promising path towards a modular and scalable quantum computer.

To control these processors (QPUs) and run advanced simulations, Quandela develops Perceval, a reference open-source SDK.

Note

Refer to the Quandela processors information page for more details.

How to access Quandela via Perceval

Before you start

To complete the actions presented below, you must have:

A Scaleway account with a valid Project ID.
A Scaleway API Key (Secret Key).
Python and Quandela Perceval installed on your local machine

Create a file with the following script to create a session, define a simple circuit, and run it. Replace $SCW_PROJECT_ID and $SCW_SECRET_KEY with your Scaleway Project ID and secret key.

Tip

Scaleway Quantum Computing provides privileged access to Quandela's physical QPUs and high-performance emulators optimized for GPUs hardware. Integration is done directly via the perceval.providers.scaleway module.

import perceval as pcvl
import perceval.providers.scaleway as scw

# The session handles resource allocation and billing
session = scw.Session(
    platform_name="QPU-ASCELLA-6PQ",  # Example of a Ascella photonic platform
    project_id="$SCW_PROJECT_ID",
    token="your-scaleway-api-key"
)

try:
    # Starts the session (provisions the resource, as it is a real QPU without booking, you will be billing on circuit & shot count)
    session.start()

    # 3. Create the Remote Processor
    processor = session.build_remote_processor()

    # 4. Define the Circuit (Simple Example: Hong-Ou-Mandel Interference)
    # Using a 50:50 Beam Splitter (Hadamard)
    circuit = pcvl.Circuit(2) // pcvl.BS.H()

    processor.set_circuit(circuit)
    processor.with_input(pcvl.BasicState("|1,1>")) # Input: 1 photon in each mode
    processor.min_detected_photons_filter(2)       # Keep only 2-photon events

    # 5. Execution (Sampling)
    sampler = pcvl.algorithm.Sampler(processor, max_shots_per_call=10000)
    job = sampler.samples(1000) # Request 1000 shots

    print("Sampling results:")
    print(job['results'])

finally:
    session.stop() # Stops the session

Save the script. In this example we save it as percival.py.
Run the script.
```
python ~/percival.py
```

How to manage a session

Context Manager (with) - You can use with scw.Session(...) as session: to automatically handle starting and stopping the session, ensuring you do not forget to stop it.