Platform / Hardware Backends

Hardware Backends

QECSync is hardware-agnostic. A unified device interface abstracts over superconducting transmon, trapped-ion, and photonic architectures. Calibration data drives decoder configuration per-device.

Device abstraction layer

One interface, any topology

QECSync introduces a DeviceSpec interface that captures the properties relevant to surface-code error correction: qubit connectivity graph, native gate set, T1/T2 per qubit, two-qubit gate error rates, and measurement assignment errors.

Hardware vendors provide calibration data in their own formats. QECSync ships parsers for common calibration schemas: IBM Quantum, Quantinuum H-series, and generic QASM 3 device specification. Once parsed, the DeviceSpec is consumed identically by both the decoder and the compiler.

Support for additional hardware vendors is available on request. Integration follows a defined protocol: provide a representative calibration JSON + a description of the native gate set. QECSync engineering evaluates compatibility and delivers backend support under NDA.

Superconducting Transmon

Nearest-neighbor 2D grid topology. Square and heavy-hex lattice both supported. Typical T2 coherence 50–300 µs. Native gates: CX / CNOT, ECR.

Supported

Trapped-Ion

All-to-all connectivity. Longer coherence times (>1 s), slower gate speeds. Natively supports Mølmer-Sørensen two-qubit gate. QECSync maps plaquette measurements accordingly.

Supported

Neutral Atom

Reconfigurable 2D arrays with mid-circuit measurement. Surface code stabilizer circuits benefit from dynamic qubit rearrangement. Backend under active development.

In development

device_init.py

from qecsync import DeviceSpec, DecoderEngine

# Parse IBM Quantum calibration export
device = DeviceSpec.from_ibm_json(
    "ibm_sherbrooke_cal_20250312.json"
)

# Print device overview
print(device.qubit_count)        # 127
print(device.mean_t2_us)         # 184.3
print(device.topology)           # "heavy-hex"

# Decoder auto-tuned to this specific device
eng = DecoderEngine.from_device(device, code_distance=5)
print(eng.weight_matrix.shape)   # (49, 49)

Integration guide

From calibration data to a tuned decoder in minutes

Export your device calibration from your hardware provider's dashboard, pass it to QECSync's parser, and your decoder receives edge-weighted syndrome graphs tuned to your specific device's noise profile. No manual weight configuration required.

Hardware Integration Guide →

Integrate your hardware backend

Integration documentation

Step-by-step guide to connecting your device's calibration data.

Read Docs →