docs(pr-x12): apply audit corrections + ground x265/x266 lens in the public H.266/ECM/NNVC/H.267 landscape#242
Merged
Conversation
…ens in the public H.266/ECM/NNVC/H.267 landscape
Two debts closed on the PR-X12 x265/x266 plan cluster:
1. The PR-X12-docs-audit.md Tier-2 corrections, unapplied since
2026-05-22, are now executed:
- blasgraph restored as the bit-exact canonical kernel home; the
fabricated symbols `blasgraph::tropical_gemm` and
`bgz17::scalar_sparse::tropical_spmv` removed — the only shipped
min-plus is the method `bgz17::ScalarCsr::spmv_min_plus`, a lossy
sibling adapter, never the canon (findings #1-#4)
- `blas_level2::batched_ssd_search` marked [PLANNED] at every
citation site (finding #5)
- per-arch DCT crossover constants tagged [UNCALIBRATED ESTIMATES];
the self-fabricating "Graviton=128" withdrawn (findings #6-#7)
- "132,710 CTUs/frame" mislabel fixed to leaves-at-8x8 (finding #8)
- false signature_kernel_pde Goursat-PDE bug claim withdrawn at all
four sites — its convergence tests pass (finding #9)
- falsifiability matrix tagged FORWARD-CONDITIONAL (finding #19)
- Tier-1 docs (woa-multiarch-orchestration, bgz-jc-substrate-
synergies) quarantined with do-not-cite headers pending rewrite
- audit doc gains a §8 applied-corrections addendum; notes that
src/hpc/codec/ now ships ans.rs (A7) + rdo.rs (A6)
2. New research doc pr-x12-h266-h267-standards-landscape.md anchors the
lens line to the real JVET trajectory, with sources: H.266/VVC
(2020, ~40-50% over HEVC, decoder 1.5-2x / encoder ~10x), ECM-16.1
(~27% over VTM-11 RA, complexity flagged impractical), NNVC v7
(NN in-loop ~9% RA each — the public antithesis of the anti-neural
rule), H.267 (>=40% over VVC Main 10 at 4K+; CfP Jul 2026 ->
submissions Nov 2026 -> evaluation Jan 2027 -> finalize ~2028).
The x266 lens doc gains a §12 reality-check addendum + F-3b
falsifier and a naming anchor ("x266" = PR-X12 3DGS scene codec,
never H.266/VVC); capstone + 3DGS plan index gain standards-watch
sections; blackboard entry appended.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K3RyLEbuNSHxxB3NTTrGki
Contributor
|
Warning Review limit reached
Next review available in: 16 minutes Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available. How can I continue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews. How do review limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please refer docs for additional details. Review details⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: ⛔ Files ignored due to path filters (1)
📒 Files selected for processing (15)
📝 WalkthroughWalkthroughThe PR updates PR-X12 documentation with audit corrections, quarantine notices, corrected kernel/API status, revised estimates and claims, and a new H.266/H.267 standards landscape linked from planning documents. ChangesPR-X12 documentation corrections
Estimated code review effort: 2 (Simple) | ~15 minutes Possibly related PRs
Suggested reviewers: Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Bugbot couldn't run - usage limit reachedBugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit. A user or team admin can review and increase usage limits in the Cursor dashboard. (requestId: serverGenReqId_a57803ec-257c-4f57-a1f8-a6ce60b833d5) |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 061cec2f82
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| 1. **Concrete trait signatures** — R-1 (`Basis<T>` + `LinearReduce` split), §8 surface (`PredictiveSignal`, `CurveOrder<const N>`, `RdoMetric`) | ||
| 2. **Quantified budgets** — R-3 LoC envelope per sub-card / per consumer + audit rule; R-4 four Plan G thresholds; R-11 4K@60fps latency budget | ||
| 3. **Math identities** — R-6 SSD-via-VNNI (`||A||² - 2A·B + ||B||²`), R-7 tropical-GEMM partition (`O(4^d) → O(d²)`, kernel at `bgz17::scalar_sparse::tropical_spmv`) | ||
| 3. **Math identities** — R-6 SSD-via-VNNI (`||A||² - 2A·B + ||B||²`), R-7 tropical-GEMM partition (`O(4^d) → O(d²)`; kernel home is `lance-graph::blasgraph` [canonical, bit-exact]; the only shipped min-plus today is the method `bgz17::ScalarCsr::spmv_min_plus` — a sibling lossy-stack adapter, NOT the canon. Corrected 2026-07-16 per `PR-X12-docs-audit.md` #1/#2) |
There was a problem hiding this comment.
This R-7 rewrite intentionally replaces bgz17::scalar_sparse::tropical_spmv with bgz17::ScalarCsr::spmv_min_plus, but the companion assertions in tests/knowledge/test_pr_x12_knowledge_docs.py still require the old symbol and wording (test_r7_cites_tropical_spmv_kernel, the substrate “Actual kernel home” checks, and the cross-doc consistency check). Running python -m pytest tests/knowledge/test_pr_x12_knowledge_docs.py -q now fails 4 tests, so the doc correction leaves the repo's knowledge-doc test suite red until those assertions are updated with the new canonical wording.
Useful? React with 👍 / 👎.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 8
🧹 Nitpick comments (1)
.claude/knowledge/pr-x12-h266-h267-standards-landscape.md (1)
184-203: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winMap each load-bearing claim to an auditable primary source.
The
[G]/[H]labels do not identify which source supports each number, and the source list relies partly on secondary commentary and Wikipedia. Add stable JVET/ITU/MPEG/arXiv identifiers per claim; retain secondary sources only as supplementary context.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.claude/knowledge/pr-x12-h266-h267-standards-landscape.md around lines 184 - 203, Update the Sources section and its associated load-bearing claims so every quantitative, timeline, requirement, and profile assertion has an explicit auditable primary-source citation using stable JVET, ITU, MPEG, or arXiv identifiers. Replace unsupported [G]/[H] labels with claim-specific references, and retain secondary sources only as supplementary context rather than sole support.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.claude/knowledge/pr-x12-bgz-jc-substrate-synergies.md:
- Around line 15-16: Keep both quarantine notices as contiguous blockquotes by
removing the blank lines or prefixing them with “>” in
.claude/knowledge/pr-x12-bgz-jc-substrate-synergies.md lines 15-16 and
.claude/knowledge/pr-x12-woa-multiarch-orchestration.md lines 13-14.
In @.claude/knowledge/pr-x12-canon-resolutions-delta.md:
- Around line 162-166: Correct the latency table in
.claude/knowledge/pr-x12-canon-resolutions-delta.md: use 129,600 exact 8×8
leaves or 130,560 leaves from padded 64×64 accounting, and remove stale
CTU-based wording. Update .claude/PR-X12-docs-audit.md to record only the
numerically verified leaf count and recompute the per-leaf budget; update
.claude/blackboard.md to match those corrected values.
In @.claude/knowledge/pr-x12-codec-cognitive-substrate-mapping.md:
- Around line 523-525: Revise the partition-optimization paragraph around the
x265 RDO description to present the tropical-semiring GEMM as a planned,
unwritten implementation rather than an existing blasgraph capability. Make the
claimed O(4^d) to O(d²) improvement explicitly conditional on successful results
from the future A6 benchmark, while preserving the existing DAG and
batched-relaxation proposal.
In @.claude/knowledge/pr-x12-h266-h267-standards-landscape.md:
- Around line 84-98: Replace the universal claim in the PR-X12
standards-landscape discussion that no NN-in-loop codec can provide legally or
scientifically defensible reconstruction with a qualified comparison of
reproducibility and model governance; in
.claude/knowledge/pr-x12-h266-h267-standards-landscape.md lines 84-98, preserve
the contrast between frozen lookup tables and in-loop neural models while
acknowledging governed NN deployments can be reproducible. Update
.claude/knowledge/pr-x12-x266-3dgs-spacetime-upscaling.md lines 362-370 to avoid
characterizing the learned track as inherently nondeterministic, and revise
.claude/plans/PR-X12-tensor-container-expansion-capstone.md lines 45-50 so
“determinism is the moat” is qualified as a reproducibility and governance
advantage.
In @.claude/knowledge/pr-x12-substrate-canon-resolutions.md:
- Around line 1271-1280: The R-11 planned test uses the obsolete ≤210 ns/CTU
threshold instead of the corrected 125 ns/leaf budget. Update the R-11 matrix
row to measure ≤125 ns/leaf, or explicitly derive and state the equivalent 64×64
CTU threshold using 64 leaves per CTU, while preserving its forward-conditional
status.
In @.claude/knowledge/pr-x12-x265-blasgraph-gemm.md:
- Around line 181-189: Update the paragraph after the Rust API block to remove
the claim that all four primitives already exist and that the BLAS layer needs
zero new lines. State instead that the wrapper APIs still require
implementation, consistent with the status note and existing building blocks.
In @.claude/knowledge/pr-x12-x266-3dgs-spacetime-upscaling.md:
- Around line 3-9: Update the naming-and-standards anchor to describe H.267 as a
prospective beyond-VVC standardization effort rather than an existing standard,
while retaining the explicit JVET source/status qualifier and the existing date
context.
In @.claude/plans/PR-X12-tensor-container-expansion-capstone.md:
- Around line 35-39: Update the H.266/VVC entry to distinguish subjective
compression savings from objective BD-rate results, using the landscape
document’s metrics (for example, up to ~50% subjective and ~31–40% objective)
instead of the combined “~40–50% over HEVC” wording.
---
Nitpick comments:
In @.claude/knowledge/pr-x12-h266-h267-standards-landscape.md:
- Around line 184-203: Update the Sources section and its associated
load-bearing claims so every quantitative, timeline, requirement, and profile
assertion has an explicit auditable primary-source citation using stable JVET,
ITU, MPEG, or arXiv identifiers. Replace unsupported [G]/[H] labels with
claim-specific references, and retain secondary sources only as supplementary
context rather than sole support.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro Plus
Run ID: 1eca357c-0cb8-4807-8554-0e1778a23f16
📒 Files selected for processing (12)
.claude/PR-X12-docs-audit.md.claude/blackboard.md.claude/knowledge/pr-x12-bgz-jc-substrate-synergies.md.claude/knowledge/pr-x12-canon-resolutions-delta.md.claude/knowledge/pr-x12-codec-cognitive-substrate-mapping.md.claude/knowledge/pr-x12-h266-h267-standards-landscape.md.claude/knowledge/pr-x12-substrate-canon-resolutions.md.claude/knowledge/pr-x12-woa-multiarch-orchestration.md.claude/knowledge/pr-x12-x265-blasgraph-gemm.md.claude/knowledge/pr-x12-x266-3dgs-spacetime-upscaling.md.claude/plans/3DGS-PLAN-INDEX.md.claude/plans/PR-X12-tensor-container-expansion-capstone.md
Comment on lines
+15
to
+16
|
|
||
|
|
Contributor
There was a problem hiding this comment.
📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
Keep both quarantine notices as contiguous blockquotes.
.claude/knowledge/pr-x12-bgz-jc-substrate-synergies.md#L15-L16: remove the blank lines or prefix them with>..claude/knowledge/pr-x12-woa-multiarch-orchestration.md#L13-L14: remove the blank lines or prefix them with>.
🧰 Tools
🪛 markdownlint-cli2 (0.23.0)
[warning] 15-15: Blank line inside blockquote
(MD028, no-blanks-blockquote)
[warning] 16-16: Blank line inside blockquote
(MD028, no-blanks-blockquote)
📍 Affects 2 files
.claude/knowledge/pr-x12-bgz-jc-substrate-synergies.md#L15-L16(this comment).claude/knowledge/pr-x12-woa-multiarch-orchestration.md#L13-L14
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In @.claude/knowledge/pr-x12-bgz-jc-substrate-synergies.md around lines 15 - 16,
Keep both quarantine notices as contiguous blockquotes by removing the blank
lines or prefixing them with “>” in
.claude/knowledge/pr-x12-bgz-jc-substrate-synergies.md lines 15-16 and
.claude/knowledge/pr-x12-woa-multiarch-orchestration.md lines 13-14.
Source: Linters/SAST tools
- tests/knowledge/test_pr_x12_knowledge_docs.py: the four assertions pinning the audit-flagged wording (tropical_spmv free function, 'Actual kernel home (current)' inversion) now assert the corrected canon instead — blasgraph canonical (kernel unwritten), shipped min-plus = the method bgz17::ScalarCsr::spmv_min_plus. Suite green (117 tests). [Codex P2] - Leaf-count arithmetic corrected everywhere: 129,600 exact 8x8 leaves at 4K (padded 130,560; the old 132,710 was ungrounded), ~129 ns/leaf budget, RDO paragraphs re-based on ~2,040 CTUs/frame, R-11 matrix row re-unitized to <=210 ns/leaf. [CodeRabbit] - Determinism claims qualified: standardized NN-in-loop decode can be bit-exact (fixed-point conformance), so the moat is stated as reproducibility-by-construction + governance, not 'NN = nondeterministic'; un-standardized AI-upscaler drift claim retained. [CodeRabbit] - H.267 described as prospective standardization effort; VVC metric split (up to ~50% subjective / ~31-40% objective BD-rate); GEMM-lens 'all four already exist' leftover removed; tropical-GEMM speedup made conditional on the A6 bench; MD028 blockquote fix in the two quarantine headers; residual affirmative stale symbols corrected in pr-x12-cam-pq-sigker-dn-tree-substrate-bindings.md. [CodeRabbit] Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01K3RyLEbuNSHxxB3NTTrGki
…pu synergy matrix The former "x266" placeholder (PR-X12 3DGS scene codec) is internally codenamed H.268 — INTERNAL ONLY, never an ITU designation (H.267 itself is still a prospective effort). Registered in the x266 lens header, the standards-landscape doc, the tensor-container capstone, and the 3DGS plan index; filenames keep the historical slug. New: .claude/knowledge/pr-x12-h268-morton-wgpu-synergies.md — the "industry-deemed-impractical vs realistically-achievable" matrix. Every claim was adversarially verified against shipped code (15-agent harvest+verify run, file:line receipts, PR-X12-docs-audit discipline): - 1x FEASIBLE-NOW: the scoping row (we do not race ECM/H.267 BD-rate; no browser encode; WebGL2 prefix-scan limits acknowledged). - 2x NEEDS-PROBE: 256x256 palette LUTs are texture-isomorphic today (dense u16, R16Uint-ready; PROBE-GPU-LUT named). - 7x OVERCLAIM-CORRECTED, with the honest wording carried: the shipped ctu.rs is an arena tree (no Morton in the codec dir; flat Morton SoA = PROBE-MORTON-CTU); D-PHASE/D-WHP are [H] with unrun probes and the J2 dither-only kill condition; the cascade certificates gate RENDER not decode; the wasm SIMD128 lane is real + CI-parity-verified but the codec is scalar; a2ui-paint's wgpu is an untested quad demo with WebGL2 unwired; rANS is single-state (CTU-granular parallelism only; PROBE-RANS-INTERLEAVE named); WGSL has no u8 and saturating add is non-associative, so GPU bit-exactness needs pinned accumulation order and excludes float EWA. Probe queue with pass/kill conditions established. Blackboard entry appended; knowledge-doc test suite green (117 tests). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01K3RyLEbuNSHxxB3NTTrGki
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Docs-only PR on the PR-X12 x265/x266 plan cluster. Two debts closed:
1. The 2026-05-22
PR-X12-docs-audit.mdcorrections — finally appliedThe audit was committed with "no edits made to the audited docs" and sat unapplied; every flagged fabrication was still live. This PR executes the Tier-2 recommendations:
lance-graph::blasgraph::tropical_gemmandbgz17::scalar_sparse::tropical_spmvare removed from all citation sites; the only shipped min-plus is correctly named as the methodbgz17::ScalarCsr::spmv_min_plusand labeled a lossy sibling adapter, never the canon.blas_level2::batched_ssd_searchmarked[PLANNED]at every site — the symbol does not exist (finding HPC backends: matrixmultiply GEMM, FFI backends, PackedDatabase cascade #5, re-verified against today's tree).[UNCALIBRATED ESTIMATES]; the self-fabricating "Graviton=128" withdrawn (findings Fix 6 clippy warnings in AVX-512 microkernels #6–CI tests #7).signature_kernel_pdeGoursat-PDE bug claim withdrawn at all four sites — the function's own convergence tests toI₀(2·√⟨u,v⟩)pass (finding feat: port learning ops, project_all, bf16_from_projections from rust… #9).pr-x12-woa-multiarch-orchestration.md,pr-x12-bgz-jc-substrate-synergies.md); full rewrites remain open work.src/hpc/codec/now shipsans.rs(A7 rANS) andrdo.rs(A6 λ-RDO), retiring audit-era debts D-CODEC-2/-3 at the code level.2. New research doc:
pr-x12-h266-h267-standards-landscape.mdGrounds the x265/x266 lens line in the real JVET/MPEG trajectory, with sources:
The x266 lens doc gains a §12 reality-check addendum + a new F-3b falsifier and a naming anchor ("x266" = the PR-X12 3DGS scene codec, never H.266/VVC). The tensor-container capstone and 3DGS plan index gain standards-watch sections. Blackboard entry appended per agent protocol.
Sibling PR in
AdaWorldAPI/lance-graph(same branch name) carries the capstone standards-watch section + the board EPIPHANIES entry.Verification
Docs-only — no code paths touched; no cargo gates required. All corrections cross-checked against the current source tree (
src/hpc/codec/contents,blas_level2.rssurface,bgz17/src/scalar_sparse.rs, blasgraph exports).🤖 Generated with Claude Code
https://claude.ai/code/session_01K3RyLEbuNSHxxB3NTTrGki
Generated by Claude Code
Summary by CodeRabbit