HITME native_mill — ONE-PAGE FROZEN TEST PROTOCOL (Rung 2 witness)
version=2026-08-24T01:10:00Z
DOI=10.5281/zenodo.22073684
mirror=https://gotokell.hitme.dev/method_rung1/

═══════════════════════════════════════════════════════════════
1. CLAIM (single · falsifiable)
═══════════════════════════════════════════════════════════════
On a laboratory-owned Linux host (or agreed witness host), using the frozen
method pack attached / linked below, transfer X unique uncompressed bytes in
Y seconds such that:
  - output SHA256 hashes match input hashes (integrity)
  - reported throughput = X/Y in GB/s (and MT/s if lane labeled)
Optional if instruments available: joules consumed (Z) and temperature
change A ± U °C at stated sensor placement.

NOT IN THIS PROTOCOL: optical Pbps/Tbps · Tokyo–London µs · absolute ΔT=0
· oncology · Tempus · Fleet effective Pbps appendix (separate doc only).

═══════════════════════════════════════════════════════════════
2. METHOD REFERENCE
═══════════════════════════════════════════════════════════════
  NATIVE_MILL_METHOD.txt (section 0 peer surface: memcpy + pack only)
  best_numbers.json — BEST vs LATEST labeled separately
  Reproduce: three runs minimum · publish median and BEST · UTC stamp

Door reference (non-binding): memcpy BEST 259093 MT/s · 253.018 GB/s
                                 pack BEST   360800 MT/s · 352.340 GB/s
Lab measures independently — not required to match; must report what was seen.

═══════════════════════════════════════════════════════════════
3. EXECUTION STEPS (lab or witness)
═══════════════════════════════════════════════════════════════
  A. Record host: model, CPU, RAM, kernel, clock source, cache policy if known
  B. Load frozen pack from USB or DOI files · verify SHA256SUMS.txt
  C. Run memcpy lane per method (or agreed equivalent dd/memcpy benchmark
     with same byte count and hash check — label class=LAB_EQUIVALENT)
  D. Run pack lane if runner available; else document why omitted
  E. Three cold-start runs · label BEST and LATEST separately
  F. Fill WITNESS_PARAGRAPH_TEMPLATE.txt · return to brian@hitme.dev

═══════════════════════════════════════════════════════════════
4. DELIVERABLES (required in report)
═══════════════════════════════════════════════════════════════
  [ ] Written statement: ISO 17025 accredited scope OR equivalent OR custom eng
  [ ] Measurement uncertainty where applicable
  [ ] Raw host logs / instrument data + chain-of-custody
  [ ] One signed witness paragraph (name · affiliation · date)
  [ ] No guarantee-of-pass language · no co-sign of parked Fleet labels

═══════════════════════════════════════════════════════════════
5. BUDGET LAW
═══════════════════════════════════════════════════════════════
  Hard cap $15,000 total program
  Phase 1 feasibility/plan: ≤ $1,500 (authorize only after this response)
  Phase 2 execute+report: ≤ $9,000
  Reserve rerun: ≥ $4,500 held until SOW names exact claim + instruments

═══════════════════════════════════════════════════════════════
CONTACT
  Brian King · Hitme · Santa Paula, CA 93060
  (805) 751-6587 · brian@hitme.dev · fleet@hitme.dev