Skip to content

Set backend parameters

Depending on the backend, some settings affecting the backend's noise characteristics can be configured.

Available parameters

Summary

Model average_nb_photons kappa_1 kappa_2 distance Availability
QPU:1Q:BOSON_4A ✅ ❌ ❌ ❌ Remote only
QPU:1Q:BOSON_4B ✅ ❌ ❌ ❌ Remote only
QPU:1Q:BOSON_4C ✅ ❌ ❌ ❌ Remote only
EMU:1Q:LESCANNE_2020 ✅ ❌ ❌ ❌ Remote & Local
EMU:6Q:PHYSICAL_CATS ✅ ✅ ✅ ❌ Remote & Local
EMU:40Q:PHYSICAL_CATS ✅ ✅ ✅ ❌ Remote & Local
EMU:15Q:LOGICAL_EARLY ✅ ✅ ✅ ✅ Remote & Local
EMU:40Q:LOGICAL_TARGET ✅ ✅ ✅ ✅ Remote & Local

Average number of photons

  • average_nb_photons is the number photons trapped in the qubit’s cavity.
  • Increasing it will exponentially decrease the number of bit-flips, at the cost of a linear increase of the number of phase-flips.
  • This parameter is available in all backends.

\(\kappa_1\) and \(\kappa_2\)

\(\kappa_1\)

  • kappa_1 is the one-photon loss rate, expressed in Hz.
  • Qubit quality decreases as kappa_1 increases, since one-photon losses causes phase-flips.
  • Boson 4 chips feature kappa_1 = 3.14*19_900.
  • This parameter is not available on a real chip and on digital twin emulators.

\(\kappa_2\)

  • kappa_2 is the two photon-loss rate, expressed in Hz.
  • Qubit quality increases as kappa_2 increases, since the exchange of pairs of photons is used to stabilize the qubit.
  • Boson 4 chips feature kappa_2 = 3.14*250_000.
  • This parameter is not available on a real chip and on digital twin emulators.

\(\kappa_1/\kappa_2\)

  • The \(\kappa_1/\kappa_2\) ratio is a good proxy for the quality of the physical chip.
  • The lower this ratio, the higher the quality of the chip.
  • Current Alice & Bob chips feature \(\kappa_1/\kappa_2 < 10^{-2}\)
  • Getting error correction to work reliably requires \(\kappa_1/\kappa_2 < 10^{-3}\)
  • Creating 100 logical qubits at a \(10^{-8}\) error rate with 1500 qubits requires \(\kappa_1/\kappa_2 < 10^{-4}\), as shown in this article
  • Running Shor's algorithm on a 2048-bit number requires \(\kappa_1/\kappa_2 < 10^{-5}\), as shown in this article

Repetition code distance

  • distance is the distance of the error correction code, i.e. the number of physical qubits used to create a logical qubits.
  • Phase-flips are exponentially removed as the distance of the code is increased, but bit-flips increase linearly.
  • This parameter is only available for logical backends.

Set parameters

Parameters may be set while initializing a backend:

from qiskit_alice_bob_provider import AliceBobLocalProvider
from qiskit import QuantumCircuit

provider = AliceBobLocalProvider()

backend = local.get_backend('EMU:1Q:LESCANNE_2020', average_nb_photons=3)
backend = local.get_backend('EMU:6Q:PHYSICAL_CATS', average_nb_photons=5)
backend = local.get_backend('EMU:40Q:PHYSICAL_CATS', kappa_2=500_000)
backend = local.get_backend('EMU:15Q:LOGICAL_EARLY', distance=13)

circuit = QuantumCircuit(...)
# ...

backend.run(circuit, shots=3000)

Or when executing a circuit:

from qiskit_alice_bob_provider import AliceBobRemoteProvider
from qiskit import QuantumCircuit

# Replace the placeholder with your actual API key in the line below
remote = AliceBobRemoteProvider(api_key='YOUR_API_KEY')

backend = remote.get_backend('EMU:1Q:LESCANNE_2020')

circuit = QuantumCircuit(...)
# ...

backend.run(circuit, shots=3000, average_nb_photons=3)