Scaleway Quantum as a Service¶

Scaleway Quantum as a Service is a managed cloud service providing on-demand access to quantum computing resources, including quantum processing units (QPUs) and quantum emulators, through a unified API.

Scaleway QaaS allows users to submit quantum workloads programmatically and integrate them into hybrid classical–quantum workflows. The service is designed to be used either directly through the Scaleway APIs or via higher-level SDK and frameworks such as CUDA-Q.

To get started, users must have an active Scaleway account and a project with Quantum Computing enabled. See the Quickstart for step-by-step instructions.

Additional information can be found in the Scaleway Quantum Computing documentation.

Setting Credentials¶

export SCW_SECRET_KEY="<secret_key>"
export SCW_PROJECT_ID="<project_id>"

Submitting¶

Python

The target backend for quantum kernel execution can be selected using the cudaq.set_target() function.

import cudaq
# Use credentials from environment variables
cudaq.set_target("scaleway")

# You can specify manually your credentials
cudaq.set_target("scaleway", project_id="<project_id>", secret_key="<secret_key>")

By default, kernels are executed on a Scaleway quantum emulator EMU-CUDAQ-H100.

To select a specific Scaleway device, set the machine argument:

machine = "EMU-CUDAQ-H100"
# machine = EMU-AER-H100 # Access to Aer emulator
# machine = QPU-EMERALD-54PQ # Access to IQM QPUs (Garnet, Sirius, Emerald)
# machine = QPU-IBEX-12PQ # Access to AQT IBEX-Q1 QPU
# machine = EMU-IBEX-12PQ-L4 # Access to AQT IBEX-Q1 emulator
cudaq.set_target("scaleway", machine=machine)

where EMU-CUDAQ-H100 identifies an emulator or QPU offer available through Scaleway Quantum as a Service. Available emulators and QPUs are listed on the Scaleway Quantum-as-a-Service webpage

The service will dynamically allocate a dedicated GPU server for your need.

This allocation can take up to few minutes. To use the same session between different script execution or users, you can specify a deduplication_id.

The session will be created if doesn’t exist, else it will retrieve and use the matching one.

machine = "EMU-CUDAQ-H100"
# The deduplication identifier is a convenient way to keep using the same resource
# between script calls or user
# Notes: The target ``machine`` must be the same as well as Scaleway project id
cudaq.set_target("scaleway", machine=machine, deduplication_id="my-workshop")

You can specify the maximal duration or the maximal idle duration to limit the billing.

machine = "EMU-CUDAQ-H100"
# The underlying QPU session will be killed after 30 minutes
# or after 5 idle minutes without new jobs
cudaq.set_target("scaleway", machine=machine, max_duration="30m", max_idle_duration="5m")

The number of shots for a kernel execution can be specified via the shots_count argument to cudaq.sample. The default value is 1000.

result = cudaq.sample(kernel, shots_count=100)

Refer to the Scaleway Quantum Computing documentation for details on device capabilities, execution limits, and pricing.

C++

To target quantum kernel code for execution on Scaleway QaaS, pass the --target scaleway flag to the nvq++ compiler.

By default, jobs are submitted to a Scaleway-managed quantum simulator.

nvq++ --target scaleway src.cpp

To execute kernels on a specific Scaleway quantum device, pass the --scaleway-machine flag to nvq++ and specify the device identifier.

nvq++ --target scaleway --scaleway-machine "<offer_name>" src.cpp

where <offer_name> refers to a Scaleway simulator or QPU available in your project. Available emulators and QPUs are listed on the Scaleway Quantum-as-a-Service webpage

Manage your QPU session¶

If you want to manually shutdown a QPU session, you can either go to the Scaleway console or do it by calling the Scaleway’s QaaS API:

# List active sessions
curl -X GET -H "X-Auth-Token: $SCW_SECRET_KEY" "https://api.scaleway.com/qaas/v1alpha1/sessions?project_id=<project_id>"

# Terminate the session
curl -X POST -H "X-Auth-Token: $SCW_SECRET_KEY" -H "Content-Type: application/json" "https://api.scaleway.com/qaas/v1alpha1/sessions/{session_id}/terminate"