Skip to content

First Hardware Data

Jump to bottom
Changhao Li edited this page May 21, 2025 · 40 revisions

Quantum Volume (QV)

Consult this page for further information on how we define quantum volume.

quantum_volume.input.json

{
  "benchmark_name": "Quantum Volume",
  "num_qubits": 4,
  "shots": 1000,
  "trials": 2,
  "confidence_level": 0.95
}

Commit: 0fe3783

Command:

python metriq_gym/run.py dispatch quantum_volume.input.json --provider <provider> --device <device>
provider device type timestamp output
ibm ibm_brussels Quantum Volume 2025-03-06T13:09:28.818993 QuantumVolumeResult(num_qubits=4, confidence_pass=True, xeb=0.886970796512533, hog_prob=0.8, hog_pass=True, p_value=0.0, trials=2)
ibm ibm_strasbourg Quantum Volume 2025-03-06T13:10:22.988862 QuantumVolumeResult(num_qubits=4, confidence_pass=True, xeb=0.7935890724559819, hog_prob=0.8005, hog_pass=True, p_value=0.0, trials=2)
ibm ibm_sherbrooke Quantum Volume 2025-03-06T13:09:54.316455 QuantumVolumeResult(num_qubits=4, confidence_pass=True, xeb=0.8802381283386798, hog_prob=0.832, hog_pass=True, p_value=0.0, trials=2)
aws arn:aws:braket:eu-north-1::device/qpu/iqm/Garnet Quantum Volume 2025-03-06T12:36:10.980330 QuantumVolumeResult(num_qubits=4, confidence_pass=True, xeb=0.7639047976125721, hog_prob=0.763, hog_pass=True, p_value=0.0, trials=2)
aws arn:aws:braket:us-west-1::device/qpu/rigetti/Ankaa-3 Quantum Volume 2025-03-06T12:34:46.668602 QuantumVolumeResult(num_qubits=4, confidence_pass=True, xeb=0.19129354830311393, hog_prob=0.6505000000000001, hog_pass=False, p_value=0.0, trials=2)

Bell State Effective Qubits (BSEQ)

bseq.input.json

{
    "benchmark_name": "BSEQ",
    "shots": 1000
}

Command:

python metriq_gym/run.py dispatch bseq.input.json --provider <provider> --device <device>
provider device type timestamp commit output
ibm ibm_brussels BSEQ 2025-02-26T21:07:07.650854 feb06fd BSEQResult(largest_connected_size=52, fraction_connected=0.4094488188976378)
ibm ibm_strasbourg BSEQ 2025-02-26T21:12:42.886197 feb06fd BSEQResult(largest_connected_size=37, fraction_connected=0.29133858267716534)
ibm ibm_sherbrooke BSEQ 2025-02-26T21:26:20.717108 feb06fd BSEQResult(largest_connected_size=100, fraction_connected=0.7874015748031497)
ibm ibm_fez BSEQ c749d44
ibm ibm_torino BSEQ 2025-03-17T15:56:30.779962 c749d44 BSEQResult(largest_connected_size=71, fraction_connected=0.5338345864661654)
ibm ibm_marrakesh BSEQ 2025-03-17T15:58:09.022280 c749d44 BSEQResult(largest_connected_size=150, fraction_connected=0.9615384615384616)
aws arn:aws:braket:eu-north-1::device/qpu/iqm/Garnet BSEQ 2025-02-26T21:21:58.614002 feb06fd BSEQResult(largest_connected_size=20, fraction_connected=1.0)
aws arn:aws:braket:us-west-1::device/qpu/rigetti/Ankaa-3 BSEQ 2025-02-26T21:24:00.500747 feb06fd BSEQResult(largest_connected_size=6, fraction_connected=0.07317073170731707)

Circuit Layer Operation Per Second (CLOPS)

Quantum Machine Learning (QML) Kernel circuit accuracy

Check Issue 64 for the plot of hardware results as a function of number of qubits. The table below are partial results.

qml_kernel.input.json

{
  "benchmark_name": "QML Kernel",
  "num_qubits": 20,
  "shots": 1000
}

Command:

python metriq_gym/run.py dispatch qml_kernel.input.json --provider <provider> --device <device>
Provider Device Date Qubits Shots Trials Accuracy
ibm ibm_sherbrooke 2025-03-07 20 1024 1 0.12109375
ibm ibm_kyiv 2025-03-09 20 1024 1 0.2841796875
ibm ibm_torino 2025-03-11 20 1024 1 0.4853515625

Clone this wiki locally