Submit a measurement
Submissions use public GitHub issues or pull requests and require a GitHub account. Review the privacy checklist before posting. Nothing is automatically added to the catalog.
Open a benchmark submission on GitHub. If the form is not yet available, use a blank issue with the fields below.
Measurement and review requirements
Submit a pull request to https://github.com/Nightglass-Protocol1/monzero or open an issue there with evidence. Maintainers review submissions before catalog inclusion. Remove keys, seeds, wallet addresses, worker identifiers, hostnames, IP addresses, usernames, paths and pool account details from evidence before posting. Copies may remain accessible after editing or deletion. Pull requests also expose commit author metadata; use a GitHub noreply email if appropriate.
Each catalog.json record needs:
id: unique lowercase slug; manufacturer/model; integer cores/threads.hashrate_hps: positive finite average, not a momentary peak.power_watts: positive measured watts, or null when unknown.power_basis: wall, cpu-package, or unknown. Only wall readings are ranked.os,miner,miner_version, ISOdateandmethodology(up to 200 chars).source_url: public HTTPS evidence including measurement interval, memory, thread settings, tuning, power measurement method and relevant sanitized logs.status: community initially;verified_byandverification_url: null.
A reviewer checks field validity, evidence, duplicates, personal-data removal and measurement basis. Approval of a community submission does not mean reproduction. Only a separate documented reproduction permits status verified, with a public reviewer identifier and HTTPS verification report. Keep original source evidence. Record corrections through reviewed commits; do not silently relabel estimates. No invented seed records. Rejected submissions are not added to the catalog; their GitHub issues and pull requests remain public. Runtime schema validation is in catalog.mjs; synthetic tests never populate production data.
Update website/mining/benchmarks/catalog.json and its updated_at date, which must be on or after each measurement date. IDs match [a-z0-9][a-z0-9-]{0,79}. All descriptive string fields are limited to 200 characters and must contain no control or bidirectional override characters. Hashrate is in H/s, not kH/s. Unknown power requires null watts; known power requires wall or cpu-package. No extra fields are permitted. Reproduction evidence must use a different URL from the original source. Reviewer judgement is still required to establish that the report is an independent measurement; schema validity does not prove that.
Validate before submitting
Run node utils/website/validate_benchmarks.mjs from the repository root. To check a separate candidate file, append its path. Then run node --test tests/website/benchmarks.test.mjs. No build or mining is required.
catalog.schema.json provides a JSON Schema for editor assistance. The Node validator is authoritative for cross-field rules, unique IDs and calendar dates; JSON Schema alone cannot enforce all of these. The pull-request workflow runs the same runtime validator and regression tests. Passing validation does not verify the measurement or authorize publication; evidence and reproduction claims still need maintainer review.
Evidence URLs must use explicit https://, contain no credentials, backslashes, whitespace or invisible control/format characters, and be at most 2048 characters. Normalized URLs differing only in hostname case, default port or fragment do not count as separate evidence. Descriptive text must contain no Unicode control, format, surrogate, line-separator or paragraph-separator characters. The CLI rejects catalog dates later than the current UTC date.
The workflow is a validation signal, not an enforced publication gate: repository maintainers must configure their branch protection and review policy. Do not mark this path-filtered workflow as required for every unrelated pull request.