Share Skippy stage wire byte accounting

[i386]

Summary

Share one stage-message wire byte estimator between Skippy telemetry and conditioned-wire simulation.

This PR is only about transfer accounting. It does not add activation striping, change message framing, or alter request/cache epoch telemetry.

Before

Skippy estimated stage message size in two places:

• binary_transport.rs for llama_stage.message_wire_bytes
• binary_transport/wire.rs for bandwidth/delay conditioning

Those estimates could drift. The conditioned-wire estimate also counted only a partial header plus tokens, activation, and raw bytes. It missed position sidebands, sampling metadata, chat metadata, and the full fixed header.

How It Works

StageWireMessage::estimated_wire_bytes() now accounts for:

• fixed stage wire header
• sampling config bytes
• logit bias bytes
• chat sampling metadata length prefix and bytes
• token sideband bytes
• position sideband bytes
• state import raw bytes
• activation payload bytes

Both telemetry and wire conditioning call the same helper.

Representative telemetry:

llama_stage.message_wire_bytes=52429612
skippy.activation_bytes=52428800

Why This Is Good

Large prefill activation frames are one of the places where pipeline transport cost becomes visible. Before adding any striped-frame or multi-stream transfer path, Skippy needs reliable accounting:

• telemetry should report the same byte count the simulator uses
• bandwidth simulation should include sidebands and metadata
• future striping thresholds can use one shared estimate
• benchmark reports become easier to compare because byte accounting is centralized

This is the low-risk prerequisite for later large activation-frame transfer work.

Compatibility

No wire-format change.

The helper estimates existing bytes; it does not encode or decode anything new.

Validation

• cargo fmt --all -- --check
• cargo test -p skippy-protocol --lib
• cargo test -p skippy-server --lib
• cargo clippy -p skippy-protocol --all-targets -- -D warnings
• cargo clippy -p skippy-server --all-targets -- -D warnings
• cargo check -p mesh-llm
• cargo clippy -p mesh-llm --all-targets -- -D warnings

Summary by CodeRabbit

• Tests

  • Added a unit test validating wire message size estimation across sampling, chat metadata, tokens, and payload scenarios.

• New Features

  • Exposed a message size estimation API to compute approximate serialized wire sizes for messages.

• Refactor

  • Centralized message size estimation for bandwidth-delay and write scheduling, improving consistency and reliability of transfer timing.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: e529cd1c-1d45-4f79-82e8-d2f60d0b8b29

📥 Commits

Reviewing files that changed from the base of the PR and between 0ce4f0c and 424780b.

📒 Files selected for processing (3)

• crates/skippy-protocol/src/binary/mod.rs
• crates/skippy-protocol/src/binary/types.rs
• crates/skippy-server/src/binary_transport.rs

🚧 Files skipped from review as they are similar to previous changes (3)

• crates/skippy-server/src/binary_transport.rs
• crates/skippy-protocol/src/binary/types.rs
• crates/skippy-protocol/src/binary/mod.rs

---

📝 Walkthrough

Walkthrough

This PR centralizes wire-size estimation into StageWireMessage::estimated_wire_bytes(), adds a unit test validating the calculation for a representative message, and updates server callsites to use the new method while removing local estimation helpers.

Changes

Centralize wire-size estimation in StageWireMessage

Layer / File(s)	Summary
Wire-size estimation implementation and test crates/skippy-protocol/src/binary/types.rs, crates/skippy-protocol/src/binary/mod.rs	StageWireMessage::estimated_wire_bytes() and private payload_wire_bytes() estimate on-wire size (fixed header, optional sampling/logit-bias, chat metadata length prefix + bytes, and kind-dependent payload sizing). A unit test verifies the computed estimate for a representative prefill message with sampling, two logit_bias items, chat metadata, tokens/positions, and activation bytes.
Update callsites to use centralized estimation crates/skippy-server/src/binary_transport.rs, crates/skippy-server/src/binary_transport/wire.rs	Replaces local helpers with message.estimated_wire_bytes() in telemetry (llama_stage.message_wire_bytes) and bandwidth-delay calculation (WireCondition::sleep_for); removes the deleted helper(s) and reformats an import list.

---

🎯 3 (Moderate) | ⏱️ ~20 minutes

• michaelneale

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 36.36% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title directly matches the PR's main objective: centralizing stage wire byte estimation by introducing a shared StageWireMessage::estimated_wire_bytes() method.
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.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/skippy-stage-wire-byte-accounting

---

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

[i386]

@michaelneale could you review this one when you get a chance?