Metriq Local Smoke¶
Date: 2026-05-06
This note records a no-QPU Metriq-Gym local simulator smoke run for the Metriq submission readiness gate.
Scope¶
The run used a standard Metriq-Gym benchmark definition and the local Qiskit Aer simulator. It did not use IBM Quantum hardware, did not upload a result, and did not submit an SCPN-specific benchmark to Metriq.
Environment¶
Metriq-Gym was executed from the isolated environment:
The project environment remained separate:
Benchmark Configuration¶
The temporary configuration used for the smoke run was:
{
"benchmark_name": "Bernstein-Vazirani",
"shots": 100,
"min_qubits": 2,
"max_qubits": 2,
"skip_qubits": 1,
"max_circuits": 1,
"method": 1,
"input_value": 1
}
The configuration was kept in /tmp for execution because it is a disposable
Metriq-Gym smoke input, not a scientific SCPN benchmark artefact.
Commands¶
/home/anulum/.venvs/scpn-metriq/bin/mgym job dispatch \
/tmp/scpn_metriq_bv_local_smoke_2026-05-06.json \
--provider local \
--device aer_simulator
/home/anulum/.venvs/scpn-metriq/bin/mgym job poll \
b96914ac-2e2f-461e-84a8-b61d81300fb2
Result¶
Metriq-Gym job ID:
Polled result summary:
job_type: Bernstein-Vazirani
provider: local
device: aer_simulator
simulator: true
device_qubits: 30
runtime_seconds: 0.017248485004529357
accuracy_score: 1.0 ± 0.0
score: 1.0 ± 0.0
Interpretation¶
This is a CLI and environment readiness check only. It confirms that the installed Metriq-Gym path can dispatch and poll a local simulator benchmark without QPU spend.
It does not establish:
- a Metriq upload;
- a Metriq-accepted SCPN benchmark;
- a QPU result;
- an SCPN package performance score;
- a comparison against the repository's Rust, VQE, DLA, FIM, or GPU benchmark artefacts.
Remaining Gate¶
The remaining Metriq task is one of:
- upload a valid Metriq-native benchmark result through the documented Metriq-Gym path after deliberate review; or
- propose an SCPN/Kuramoto-XY benchmark schema upstream and wait for acceptance before any SCPN-specific result upload.