decoder-comparison-pymatching-mwpm

surface-code-memory-decoder-comparison v0.1.0

Hash mismatch means recalculating the hash from the stored data does not match the original submission. This should not normally happen -- treat this result as unverified until it's investigated. It is not by itself proof of scientific correctness or independent reproduction -- see verification evidence below for that.

Results at a glance

9 metric points

note: Shared hashed samples; paired inference lives in the ketqat-benchmarks results.
points: 9
Executed: Jul 31, 2026, 2:27 AM

Run context

Domain: QEC
Schema: 0.1
SDK / runner: 0.2.0
Repository: Not recorded in the submission
Commit: Not recorded in the submission

Reproducibility

Hash mismatch

058a1101427a6e05de5c8a091dd8a65bea0f8f654713b87244228c756ef92478

Command

ketqat run experiment.yaml --output run.json

Download bundle

Cite this

@misc{ketqat_run_decodercomparisonpymatchingmwpm058a1101_2026,
  title        = {decoder-comparison-pymatching-mwpm},
  author       = {{KetQat contributors}},
  year         = {2026},
  howpublished = {\url{https://ketqat.com/runs/decoder-comparison-pymatching-mwpm-058a1101}},
  note         = {KetQat benchmark run on surface-code-memory-decoder-comparison v0.1.0. Retrieved via KetQat, a vendor-neutral registry for reproducible quantum error-correction and quantum-algorithm research.}
}

Performance

Logical error rate against physical error rate. Error bars show reported standard error when available.

d=3d=5Log y-axis

Complete result data

Measured values are shown without interpolation or threshold estimates.

Metric	Distance	Physical p	Value	Std. error	Shots	Failures	Latency ms	Runtime s
logical_error_rate	3	0.008	3.160e-2	2.47e-3	5000	158	-	-
logical_error_rate	3	0.008	3.020e-2	2.42e-3	5000	151	-	-
logical_error_rate	3	0.02	1.462e-1	5.00e-3	5000	731	-	-
logical_error_rate	3	0.02	1.308e-1	4.77e-3	5000	654	-	-
logical_error_rate	5	0.008	3.900e-2	2.74e-3	5000	195	-	-
logical_error_rate	5	0.008	3.980e-2	2.76e-3	5000	199	-	-
logical_error_rate	5	0.02	2.695e-1	9.92e-3	2000	539	-	-
logical_error_rate	5	0.02	2.545e-1	9.74e-3	2000	509	-	-
logical_error_rate	3	0.02	1.419e-1	2.47e-3	20000	2837	-	-

Verification evidence

No verification evidence recorded.

Reproduction requests

Open

requested by admin automation · 2026-08-02

Independent reproduction wanted: rotated surface code d=3, circuit-level depolarizing p=0.001, shared-Stim detector sampling, pymatching MWPM via from_detector_error_model. The run's manifest, seed and command are on the run page; the reproducibility hash pins the payload. No independent party has re-executed this yet -- platform automation regenerating its own runs does not count, and this platform refuses to attest itself.

Comment · admin automation · 2026-08-02
What a valid attempt needs: (1) re-run from the manifest with the recorded seed and package versions, (2) import the result so it has its own registry page, (3) record EVIDENCE here naming that run with a durable URL and the exact command, (4) attest. MATCHED evidence without attestation leaves this request OPEN by design: hash verification never equals reproduction.

Environment and benchmark configuration

Environment

{
  "hardware": {},
  "packages": {
    "ketqat-runner": "0.2.0"
  }
}

Configuration

{
  "seed": 42,
  "shots": 100,
  "distance": 3,
  "physical_error_rate": 0.001
}