facet: per-axis LCP reads the single register — the fold was done at mint

[AdaWorldAPI]

What

shared_prefix_tiles already reads the whole facet as one register (u128 xor + trailing_zeros/16); hi_distance/lo_distance (shared6) still gathered six strided bytes per axis into a chain and walked them in a loop.

Operator correction on the first cut (which re-folded the gathered chain into a u64): fold the {0}{1} -f logic once — and the u128 facet already holds both axes by position, so an axis prefix is the whole-facet xor masked to that axis's tier bytes (HI_BYTES = 5,7,…,15; LO_BYTES = 4,6,…,14), then trailing_zeros/16 past the classid. No gather, no per-call re-fold. const fn, no new deps.

Falsifier compares against the loop it replaced at every divergence tier on both axes plus the identical case; disable-run (swap HI_BYTES/LO_BYTES) fails at hi flip at tier 0.

Board: E-FORMAT-SLOT-FOLD-IS-THE-SAME-OP-AS-THE-VL-DESCENT-1 (EPIPHANIES prepend, same PR).

Reality check — before/after, 64K random facet pairs, release, best of 7

arm	ns/pair
hi+lo distance, byte loop (before)	12.5–14.0
hi+lo distance, gathered chain → u64 fold (first cut)	8.3–9.3
hi+lo distance, masked single-register readout (shipped)	5.8 (2.9 per axis)
prefix_distance (u128, whole facet)	2.6–3.1

0 mismatches over 65 536 pairs at every step. Both axes together now cost what one whole-facet readout costs — the compare is the 1–4-cycle op and nothing is gathered in front of it.

🤖 Generated with Claude Code

https://claude.ai/code/session_01DCfrD5y19cvFc4AoyydXYv

Summary by CodeRabbit

• Performance

  • Improved facet distance calculations, reducing measured latency from 12.5 ns to 5.8 ns.
  • Optimized handling of short vector tails, with measured speedups of up to 8× in applicable scenarios.

• Bug Fixes

  • Improved consistency when comparing identical facets and facets that differ at individual hierarchy levels.

• Documentation

  • Added documentation covering the updated facet processing approach and performance findings.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Essentials

Run ID: 8d3d820c-5656-43f8-86df-63ba3dd28563

📥 Commits

Reviewing files that changed from the base of the PR and between e054dcf and 45c5199.

📒 Files selected for processing (2)

• .claude/board/EPIPHANIES.md
• crates/lance-graph-contract/src/facet.rs

Included review availability: 0 reviews are currently available. Your included PR review attempts over the past 7 days set your current allowance at 3 reviews per hour.

---

📝 Walkthrough

Walkthrough

FacetCascade now computes axis distances with masked u128 XOR values and trailing-zero counts. Tests compare the folded results with the former loop across identical facets and tier divergences. Documentation records benchmark results and follow-up gaps.

Changes

Facet prefix folding

Layer / File(s)	Summary
Packed facet fold and validation crates/lance-graph-contract/src/facet.rs, .claude/board/EPIPHANIES.md	FacetCascade uses masked XOR folding for hi_distance and lo_distance. Tests cover identical facets and each tier divergence against loop-based results. The notes record benchmark measurements and follow-up gaps.

Priority: ⬇️ Low

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Refactor

Suggested reviewers: claude

Merge Risk: ⚪ Minimal · up to 45c51

The facet distance optimization preserves axis prefix calculations, with no actionable merge risk identified.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 77.78% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 9 functions across 1 files. (1 skipped: 1…	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly describes the main change: per-axis LCP reads the pre-folded single-register facet value.

Full details: Docstring Coverage

Explanation

Docstring coverage is 77.78% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 9 functions across 1 files. (1 skipped: 1 unsupported.)

• Fix all pre-merge checks with AI

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

• Create stacked PR
• Commit on current branch

---

_{A rabbit reads each line,
The patch grows clear beneath the moon,
Small changes hop in place,
Tests guard the garden path,
Reviews bloom before the dawn.}

_{Comment @coderabbitai help to get the list of available commands.}

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot 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_c76396a6-84a2-402b-aea1-3661b180a419)