Backends

Overview

A backend is the execution target for your quantum circuits. You obtain a backend from a QuantumRingsProvider and use it to run jobs.

The backend you choose determines whether execution runs on:

  • CPU only

  • GPU

  • Hybrid CPU + GPU

Backend options

The SDK provides three backend names:

scarlet_quantum_rings — CPU

CPU-only execution.

Use this backend if:

  • You installed the CPU-only package

  • You do not have an NVIDIA GPU

  • Your circuits are small (few qubits and shallow depth)

For small circuits, CPU execution is often faster than GPU.

amber_quantum_rings — GPU

Full GPU execution.

Use this backend if:

  • You installed the GPU-enabled package

  • You have a supported NVIDIA GPU

  • Your circuits are larger or deeper

serin_quantum_rings — Hybrid

Hybrid execution (CPU with GPU offload).

Use this backend if:

  • You installed the GPU-enabled package

  • Your GPU has limited memory

  • You want GPU acceleration for large operations but CPU for the rest

Selecting a backend

Core SDK example:

from QuantumRingsLib import QuantumRingsProvider

provider = QuantumRingsProvider()
backend = provider.get_backend("scarlet_quantum_rings")

If you have saved your backend in the local configuration file, you may omit the name:

provider = QuantumRingsProvider()
backend = provider.get_backend()

When using the Qiskit toolkit, the backend name is provided when creating the runtime service or backend. See Toolkit for Qiskit.

Common guidance

  • CPU-only installation → use scarlet_quantum_rings.

  • GPU-enabled installation → use amber_quantum_rings or serin_quantum_rings.

  • Few qubits and shallow circuits → scarlet_quantum_rings.

  • Large or deep circuits → amber_quantum_rings.

  • Limited GPU memory → serin_quantum_rings.

For precision and runtime options, see Run Settings.

See also