Skip to content

docs(triangle-protocol): public L0-L4 measurement spec#27

Open
gHashTag wants to merge 2 commits into
mainfrom
feat/triangle-protocol
Open

docs(triangle-protocol): public L0-L4 measurement spec#27
gHashTag wants to merge 2 commits into
mainfrom
feat/triangle-protocol

Conversation

@gHashTag

@gHashTag gHashTag commented Jul 4, 2026

Copy link
Copy Markdown
Owner

Reproducibility-first artifact — the operational expression of the auditability axiom in the δ paper (#24, merged). This is the artifact we ask incumbent vendors to produce and do not; we produce it for ourselves first.

What it defines

  • Hardware under test: P201Mini ×3 (Zynq-7020 + AD9361), with the real 2026-07-01 baselines already in repo — AD9361 5.8 GHz loopback at 108.6 dB SNR (radio/README.md), M1 crypto RC=0 on-device (smoke/M1_RESULTS.md).
  • Closed-set honesty tags: hw / sim / openwifi-baseline / raw-IQ-ceiling / ?. A number without a tag is a defect in this document, not a result.
  • Levels L0–L4: exact commands, capture files, pass criteria, output schema (which benchmark cell each fills).
  • L3 honesty fork: openwifi-baseline and raw-IQ-ceiling published as two columns, never one masquerading as the other — because Tri-Net has no in-house OFDM PHY yet and won't pretend otherwise.

What it deliberately does NOT do

  • Does not promise a throughput number. Promises a method.
  • Does not hide prerequisites — §10 executability matrix names every blocker.

Honest status (verified this session)

Nothing in L0–L4 is runnable this session: pzp201mini is not resolvable from the dev host (nodes offline / not on network). The protocol's value today is that it exists publicly and defines exactly what runs the moment the hardware comes back — which is the property the δ paper demands of vendors and which we therefore demand of ourselves.

Sequencing

This PR is the spec. Follow-up work (when nodes are online): bin/microbench-m1 (L1), then capture files smoke/TRIANGLE_SANITY.md (L0), smoke/L1_MICROBENCH.md, etc. Each becomes its own PR against this protocol.

Anchor: φ² + φ⁻² = 3

SSD DDD added 2 commits July 4, 2026 12:36
…ty-first)

Defines, in advance and in public, exactly how Tri-Net measures its own
MANET radio performance on P201Mini hardware, so every published number is
reproducible by a third party. Operational expression of the auditability
axiom in the delta paper (docs/WAVE_N3_AUDITABILITY_GAP_2026-07-04.md):
the existence of a reproducible protocol is the primary quantity, not the
throughput number. This is the artifact we ask incumbent vendors to produce
and do not; we produce it for ourselves first.

Contents:
  - Hardware under test (P201Mini x3, Zynq-7020 + AD9361) with real 2026-07-01
    baselines: AD9361 5.8 GHz loopback 108.6 dB SNR, M1 crypto RC=0 on-device.
  - Closed-set honesty tags: hw / sim / openwifi-baseline / raw-IQ-ceiling / ?.
    A number without a tag is a defect, not a result.
  - Levels L0-L4 with exact commands, capture files, pass criteria.
  - L3 honesty fork: openwifi-baseline vs raw-IQ-ceiling published as TWO
    columns, never one masquerading as the other.
  - Executability matrix: NOTHING runnable this session (pzp201mini not
    resolvable from dev host) -- blockers named, not hidden.
  - Reproducibility checklist for a third party.

No numbers promised. A method promised. Numbers follow when hardware does.
docs/TRIANGLE_MEASUREMENT_PROTOCOL.md L0 and the M-map both reference
'docs/LOCAL_FLASH.md section 3' as the M1x3 procedure, but the file did not
exist on main. This commit adds it.

Derived entirely from the already-verified 2026-07-01 single-board results in
smoke/M1_RESULTS.md (M1 graduation) and radio/README.md (AD9361 loopback) --
nothing new, just the existing one-board procedure parameterized for three.

Contents:
  - Prereqs (incl. PL/FPGA NOT flashed as of 2026-07-01 -- does not block M1
    on PS, does block M2+ PL modem; flagged explicitly).
  - Host toolchain (rustup armv7-unknown-linux-musleabihf, record sha256).
  - Per-node procedure: scp smoke-m1, run RC=0, uname/hostname/iio ad9361,
    ad9361_loopback.sh + analyze_tone.py (peak within +-10 Hz, SNR >= 100 dB).
  - M1x3 graduation matrix (3 rows appended to smoke/M1_RESULTS.md).
  - Output: smoke/TRIANGLE_SANITY.md (L0 capture file).
  - Honesty: PS-only (no RF range/mesh claim), LOOPBACK=1 on-chip not radiated.
  - Executability: not runnable this session (nodes unreachable) -- hardware gap,
    not code gap.

Closes the 'references a non-existent doc' gap so M1x3 is ready-to-run the
moment a node is powered and on the network.
@gHashTag
gHashTag marked this pull request as ready for review July 7, 2026 12:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant