HMAC-64 envelope research: no ≥10% exact shortcut

[arkadianet]

First-principles investigation of the actual BIP39 PBKDF2-HMAC-SHA512 hot loop (4094 HMAC-64 compressions per seed). Does not change production kernels.

Result

Negative on the ≥10% bar. The envelope and the iteration boundary were expanded completely and tested. The only exact work reduction that survives is the already-known pad64 first-expand specialisation (~3% counted ops). New geometry (early digest words, DM-add absorption, inner/outer shared subgraphs, MiniHMAC algebra) is either false or ≤0.65%.

Full kill ledger and the seven required findings: docs/hmac_envelope_research.md.

What was implemented

• Exact word-level HMAC-64 model, bit-checked against production hmac_sha512
• First-expand DAG (23 add + 8 σ0 + 14 σ1)
• 160-round / iteration-boundary ready-round analysis
• MiniHMAC-8 homomorphism and f⊕f² affinity probes
• 8-bit superopt of Σ1+Ch (80k programs)
• Release timing of a Davies–Meyer absorption prototype

cargo test -p erg-vanity-crypto hmac_envelope

Intentionally not shipped

The ~3% first-expand specialisation already lives on cursor/pbkdf2-hmac64-specialized-fbd7. This cycle does not polish it into the GPU kernel.

Base: main.

  

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: b40e9604-9b99-4998-a6f3-96580a273d56

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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.}