perf(map): index dense numeric key ranges

[proggeramlug]

Summary

• retain the existing numeric-key hash index as the complete fallback
• add a bounded, base-relative range index after at least eight exact nonnegative integer keys demonstrate a dense span
• serve covered hits and misses with a bounds check plus array load, independent of the absolute key magnitude
• keep sparse, fractional, negative, tagged, and oversized ranges on the hash path
• keep dense entry offsets exact across insertion, ordered deletion, and Map.clear()

The range is admitted only when its span is at most 4x the observed integer-key count and is capped at 1,048,576 slots. A run beginning at 1,000,000 therefore costs the same as a run beginning at zero, while isolated large keys cannot trigger a magnitude-sized allocation.

Mac-mini evidence

Unchanged codehz/ecs row: 5k entities: 3 commands each + sync (15k total commands).

Exact 11-pair alternating cohort, repeat 64, two warmups and six measured rounds per process:

• control SHA-256: 03b8d566953b5ffc7915341b1c946de5941b4b1344ee7c20b9bbd761b9409208
• candidate SHA-256: e432c9d2d0124ccbc336d3458ae3539f2a3a9048658ddf92ef49231934301b44
• control median: 20.989459 ms
• candidate median: 19.551747 ms
• median paired improvement: 6.852% (range 5.907–8.061%)
• candidate wins: 11/11
• semantic process oracles: 22/22

The post-change native sample reduces map::find_key_index from 355/3,728 main-thread samples (9.52%) to 48/2,983 (1.61%), an 83.1% relative reduction in normalized share. That confirms that the measured gain comes from removing the intended hash probes.

Node 26.5.1 versus candidate, 11 alternating pairs, reports 2.882646 ms Node and 22.222847 ms Perry: 7.709x by independent medians and 7.642x by paired medians, with 22/22 semantic oracles. The preceding landed stack measured 8.172x.

Verification

• cargo test -p perry-runtime --lib --quiet: 2,706 passed, 0 failed, 4 ignored
• focused Map module: 5 passed, including dense high-base activation, sparse/fractional/negative fallback, mixed-index ordered deletion, direct-index stability, and clear
• python3 scripts/gc_runtime_root_holders.py --self-test
• python3 scripts/gc_runtime_root_holders.py
• cargo fmt --all -- --check
• git diff --check

[coderabbitai]

Warning

Review limit reached

Next included review available in 12 minutes.

View limit details

Limit details: You’ve used all 8 included reviews currently available.

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 5667cdcb-8aa3-466a-a171-dba80a6fba0d

📥 Commits

Reviewing files that changed from the base of the PR and between 84ff806 and 05e3930.

📒 Files selected for processing (2)

• changelog.d/8867-dense-numeric-map-index.md
• crates/perry-runtime/src/map.rs

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[proggeramlug]

Landed on main via the #8870 batch.