[justinchuby]

Replacement for #1210, rebased onto current main after the wiki base (#1190) landed as #1267. The original is stacked on #1190's head branch and cannot be squash-merged directly (GitHub stacked-PR 403). Content is identical to #1210's own diff: 154 files under site/, one Pages workflow, .gitignore rules, and two small edits to wiki/.

What this adds

The Quartz 5 static-site machinery that publishes the wiki/ content (added in #1190/#1267) to GitHub Pages:

• site/quartz/ — the Quartz engine, package.json + committed package-lock.json, quartz.config.yaml, plugin lock (quartz.lock.json).
• site/scripts/ — Python validators (validate_wikilinks.py, validate_site.py) and a test suite.
• .github/workflows/wiki-pages.yml — build + deploy workflow.

Practical review (the questions that matter for publishing machinery)

• Build dependency / CI job — acceptable and documented? Yes. The new wiki-pages.yml pins every action to a full SHA, runs with top-level permissions: contents: read (the deploy job separately requests pages: write / id-token: write), sets persist-credentials: false, installs via npm ci from the committed lockfile (reproducible), and deploys only on push to main (build-only on PRs/branches). It is self-contained and does not touch the Rust build or any existing workflow.
• Commits generated output that will rot? No. .gitignore excludes site/quartz/node_modules/, public/ (the built site), .quartz/, .quartz-cache/, vendored static assets, and tsconfig.tsbuildinfo. No node_modules, built HTML, or public/ is committed. The committed package-lock.json pins exact dependency versions.
• Conflicts with the existing docs layout? No. There is no pre-existing site/ or Pages/Quartz/mkdocs workflow on main; the existing workflows (ci, publish, wheels, benchmark, squad-*, …) are untouched. The site lives under site/ and reads wiki/ (quartz build -d ../../wiki); it does not move or rewrite docs/.
• Config coherence. quartz.config.yaml baseUrl: justinchuby.github.io/onnx-genai matches the /onnx-genai/ base path the post-build validators (validate_site.py public /onnx-genai/, rewrite_repository_links.py) expect.

Verification I could run

• Wikilink validation passed (validate_wikilinks.py ../../wiki): Validated 96 wikilink(s) across 20 note(s).
• Nothing in cargo or the Rust test suite references site/, so this cannot affect the crate build or tests.

Gap I must be honest about — the end-to-end site build

I could not run the actual Quartz build (npm run wiki:build) in this environment. npm ci fails because the pinned transitive dependency @xmldom/xmldom@0.8.14 (pulled via pixi.js) is not available on the corporate npm proxy (which only serves 0.9.x for that package), and the public registry is TLS-blocked here. There is no lockfile-preserving workaround from this sandbox.

Consequently the Python test suite shows 37/78 failures — but every failure is a missing-node-dependency error (Cannot find package 'acorn' …) from tests that shell out to the un-installed Node scripts, not a logic defect; the network-free checks pass.

The definitive build proof therefore rests on the wiki-pages.yml job this PR adds, which runs on ubuntu-latest against public npm where 0.8.14 exists. Blast radius is contained: a build failure would surface only in this isolated workflow / Pages deploy and cannot break the Rust build or test suite. Merging on the machinery review + link validation, with the JS build unverified-locally and delegated to CI, clearly flagged.

Closes #1210.

Co-authored-by: Copilot 223556219+Copilot@users.noreply.github.com

[github-actions]

🔴 Benchmark Regression Detected

Comparison of criterion micro-benchmarks: PR head vs merge-base, measured on the same runner in the same job (base first → PR second).

ℹ️ Absolute times are informational only — they vary with runner load. The % change column is the reliable signal because both sides ran under identical conditions.

Status	Scenario	Base	PR	Change
🔴	gather/large_f32_threads=1-internal/131072	24.46 µs	36.25 µs	+48.2%
🔴	gather/small_bf16_threads=1-internal/4096	489.5 ns	718.7 ns	+46.8%
🔴	matmul/large_generic_bf16_threads=8/32x1024x1024	2.13 ms	3.00 ms	+41.0%
🔴	matmul/medium_generic_bf16_threads=8/32x512x512	470.44 µs	627.56 µs	+33.4%
🔴	gather/medium_f32_threads=1-internal/32768	3.95 µs	5.23 µs	+32.5%
⚠️	block_quantized_matmul_cached_dense/mxfp4_preexpanded_dense_oncelock_like_proxy/1x1024x1024	55.36 µs	70.53 µs	+27.4%
⚠️	gather/large_f16_threads=1-internal/131072	13.61 µs	17.20 µs	+26.4%
⚠️	matmul/small_generic_bf16_threads=8/1x256x256	39.45 µs	49.52 µs	+25.5%
⚠️	matmul/large_generic_bf16_threads=1/32x1024x1024	2.28 ms	2.85 ms	+25.0%
⚠️	grammar_masking/llguidance_compute_mask/32	75.51 µs	93.27 µs	+23.5%
⚠️	gather/large_bf16_threads=1-internal/131072	12.68 µs	15.33 µs	+20.9%
⚠️	sampling_latency/min_p_per_token	237.73 µs	285.42 µs	+20.1%
⚠️	gather/medium_f16_threads=1-internal/32768	2.55 µs	2.98 µs	+16.8%
⚠️	matmul/large_generic_f32_threads=8/32x1024x1024	4.51 ms	5.26 ms	+16.8%
⚠️	kv_cache/alloc_dealloc_pages	39.62 µs	46.00 µs	+16.1%
⚠️	qwen3_sampling_processors/top_p_full_sort_after_top_k_baseline	3.71 ms	4.30 ms	+15.8%
✅	block_quantized_matmul_cached_dense/mxfp4_uncached_dequant_each_call/1x1024x1024	858.26 µs	962.79 µs	+12.2%
✅	matmul/large_generic_f16_threads=8/32x1024x1024	106.33 µs	118.39 µs	+11.3%
✅	sampling_latency/top_p_per_token	457.60 µs	508.61 µs	+11.1%
✅	matmul/medium_generic_f32_threads=1/32x512x512	2.56 ms	2.85 ms	+11.1%
✅	gather/small_f16_threads=1-internal/4096	575.2 ns	628.4 ns	+9.2%
✅	sampling_latency/top_k_per_token	62.56 µs	67.92 µs	+8.6%
✅	tokenization/decode_tokens_per_second	7.24 ms	7.84 ms	+8.2%
✅	tokenization/encode_tokens_per_second	467.30 µs	494.45 µs	+5.8%
✅	logit_processing/seven_processor_chain_per_step	340.00 µs	358.60 µs	+5.5%
✅	qwen3_sampling_processors/top_p_fast_after_top_k	554.20 µs	583.95 µs	+5.4%
✅	matmul/large_generic_f32_threads=1/32x1024x1024	10.11 ms	10.52 ms	+4.1%
✅	matmul/medium_generic_f16_threads=8/32x512x512	38.88 µs	40.37 µs	+3.8%
✅	sampling_latency/greedy_per_token	3.39 µs	3.49 µs	+3.2%
✅	qwen3_sampling_processors/top_k_top_p_full_sort_baseline	5.78 ms	5.95 ms	+2.8%
✅	qwen3_sampling_processors/top_k_partial_selection	163.79 µs	165.17 µs	+0.8%
✅	matmul/medium_generic_bf16_threads=1/32x512x512	579.29 µs	577.14 µs	-0.4%
✅	reduce_mean/medium_f32_threads=1-internal/65536	334.38 µs	326.58 µs	-2.3%
✅	matmul/small_generic_f32_threads=1/1x256x256	49.75 µs	47.85 µs	-3.8%
✅	matmul/small_generic_f16_threads=1/1x256x256	45.79 µs	43.96 µs	-4.0%
✅	matmul/medium_generic_f32_threads=8/32x512x512	1.39 ms	1.33 ms	-4.3%
✅	matmul/large_generic_f16_threads=1/32x1024x1024	92.76 µs	88.55 µs	-4.5%
✅	gather/small_f32_threads=1-internal/4096	888.5 ns	839.4 ns	-5.5%
✅	qwen3_sampling_processors/top_k_full_sort_baseline	2.50 ms	2.34 ms	-6.2%
✅	matmul/medium_generic_f16_threads=1/32x512x512	41.50 µs	38.79 µs	-6.5%
✅	qwen3_sampling_processors/top_k_top_p_fast	742.12 µs	690.56 µs	-6.9%
✅	gather/medium_bf16_threads=1-internal/32768	3.35 µs	3.06 µs	-8.7%
✅	reduce_mean/small_f32_threads=1-internal/4096	18.24 µs	15.90 µs	-12.8%
✅	matmul/small_generic_bf16_threads=1/1x256x256	57.03 µs	49.33 µs	-13.5%
🟢	block_quantized_moe_cached_dense/mxfp4_uncached_expert_dequant_each_call/rows=1,H=256,I=256,E=4,top_k=1	628.41 µs	521.40 µs	-17.0%
🟢	reduce_mean/large_f32_threads=1-internal/262144	1.33 ms	1.10 ms	-17.8%
🟢	add/small_bf16_threads=1-internal/1024	552.1 ns	453.2 ns	-17.9%
🟢	add/small_f16_threads=1-internal/1024	592.7 ns	485.1 ns	-18.1%
🟢	add/medium_f16_threads=1-internal/262144	142.38 µs	111.12 µs	-22.0%
🟢	add/medium_bf16_threads=1-internal/262144	133.34 µs	103.39 µs	-22.5%
🟢	add/large_f32_threads=1-internal/4194304	957.11 µs	683.03 µs	-28.6%
🟢	add/small_f32_threads=1-internal/1024	281.9 ns	198.8 ns	-29.5%
🟢	block_quantized_moe_cached_dense/mxfp4_cached_dense_expert_repeated_call/rows=1,H=256,I=256,E=4,top_k=1	142.35 µs	99.55 µs	-30.1%
🟢	matmul/small_generic_f16_threads=8/1x256x256	65.57 µs	45.44 µs	-30.7%
🟢	add/large_f16_threads=1-internal/4194304	2.42 ms	1.66 ms	-31.3%
🟢	add/medium_f32_threads=1-internal/262144	37.78 µs	25.15 µs	-33.4%
🟢	add/large_bf16_threads=1-internal/4194304	2.52 ms	1.66 ms	-34.0%
🟢	block_quantized_matmul_cached_dense/mxfp4_cached_dense_repeated_call/1x1024x1024	137.38 µs	76.59 µs	-44.2%
🟢	matmul/small_generic_f32_threads=8/1x256x256	88.26 µs	46.89 µs	-46.9%

Visual flags: ⚠️ ≥ 15% slower, 🔴 ≥ 30% slower — calibrated against measured runner noise (~27% worst-case on multi-threaded matmul)

Host info

CPU: Apple M1 (Virtual)
Cores: 3
OS: Darwin 25.5.0 arm64
Rust: rustc 1.97.1 (8bab26f4f 2026-07-14)
Load avg: { 4.21 3.79 7.48 }

What this cannot catch

• Regressions in code paths not covered by these benchmarks (e.g., end-to-end decode with a real model)
• Sub-threshold regressions that compound over multiple PRs
• Performance changes that only manifest under GPU execution
• Latency changes in the ORT integration path (these benchmarks exercise the native Rust kernels)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.96%. Comparing base (1ad71c4) to head (b1dc1e2).
⚠️ Report is 36 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##           main    #1272       +/-   ##
=========================================
+ Coverage      0   79.96%   +79.96%     
=========================================
  Files         0      361      +361     
  Lines         0   157076   +157076     
  Branches      0   157076   +157076     
=========================================
+ Hits          0   125613   +125613     
- Misses        0    26829    +26829     
- Partials      0     4634     +4634     

Flag	Coverage Δ
offline	79.96% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.
see 361 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.