I measured X. The published number is Y. Why?
Enter what you measured for a hardware × model × precision and it is compared with the published expected range from the normalized benchmark database (or the estimator's fallback when nothing was measured). You get a verdict, the gap ratio, and the likely causes ranked by fit — each with the on-device command that settles it. Give the JetPack release and power mode and the Jetson Configuration Checker runs too.
Method v1.4.0 · data verified 2026-09-28 · registries: benchmarks-v2, jetpack-config, power-modes · methodology
01 · What you measured
Advanced
Warm-up and timed iterations rank or rule out the warm-up cause only when your measurement is below the expected range; inside the range no cause is ranked.
Your board (optional — sharpens the causes, runs the Jetson Configuration Checker)
These answers raise, demote or rule out the likely causes only when your measurement is below the expected range — each ranked cause says which answer moved it, and a cause an answer excludes is listed under “Ruled out by your answers”. Inside the range no cause is ranked. Deployment is informational (no cause is ranked on it); the install method is checked by the Jetson Configuration Checker when you give the JetPack release or power mode, and on its own only an ISO install bears on a cause (the Super flash).
03 · Submit a measurement
If your run followed the measurement protocol, submit it below — it is reviewed against the protocol before it appears as a class C row with your attribution. Run the check above first and use "Submit this measurement" to prefill this form.
Submit a measurement
04 · How this works
The reference is the Benchmark Explorer's expected range
For your hardware, model, variant and precision, the reference is the min / median / max of every published row (or the estimator's B → C → D fallback when nothing was measured). Your measurement is converted into the same metric if needed (fps ↔ latency at your batch size), then gap ratio = measured ÷ median. WITHIN_RANGE sits inside the range ±15%; BELOW_RANGE / ABOVE_RANGE fall outside it (direction-aware for latency, where lower is better); NO_REFERENCE means nothing exists to compare against yet.
Causes are ranked, not guessed
Each cause has a heuristic prior — set from what you reported (power mode, cooling, clocks, warm-up and timed iterations, Super flash, install method, scope, runtime, batch, input size) — multiplied by how well its typical effect size matches your observed gap; the result is published as an order, never as a probability, and each cause says which of your answers moved it. A cause one of your answers excludes — clocks locked, warm-up discarded, the Super flash present, or a Super flash that the reference's preset does not need — is listed under “Ruled out by your answers”; a condition no cause can use on your hardware is labelled “Not used for the ranking”. Every ranked cause carries the on-device command that settles it, its expected output, and a remediation; when you give a JetPack and power mode, the Jetson Configuration Checker's non-PASS checks feed directly into the ranking.
Submitting your measurement closes the gap for the next person
If your run followed the measurement protocol, submit it — it is reviewed before it appears as a class C row with your attribution, replacing an estimate or filling a hole in the database entirely.
Full formulas, source list and what invalidates a result: methodology. Reference rows: /benchmarks/. Measurement protocol: how to submit a comparable number. Query the database directly: Benchmark Explorer.