Benchmark cold, warm, and incremental materialization CPU

[flyingrobots]

Parent: #758
Related implementation: #738

Actor story

As a git-warp maintainer, I need reproducible cold, warm, and incremental materialization measurements so that an optimization is proven against the same causal workload and later changes cannot quietly restore full replay or redundant CAS work.

Required work

• Define a versioned, deterministic causal corpus generated from a fixed seed.
• Keep corpus generation outside timed regions while reporting its logical size and topology.
• Run cold materialization, warm git-cas reuse, and one bounded suffix update as distinct scenarios.
• Capture process user/system CPU, wall time, peak RSS/heap, Git operations, git-cas evidence, and output cardinality.
• Include warmups and repeated samples; report median and dispersion.
• Execute base and head on the same runner/toolchain with order bias controlled.
• Emit a stable JSON result schema suitable for CI comparison and historical publication.

Acceptance criteria

• The corpus is deterministic and versioned.
• Cold, warm, and incremental scenarios assert identical semantic results where expected.
• Timed regions and excluded setup are documented.
• CPU and wall time are reported separately.
• Warm reuse proves that git-cas materialization state is actually hit.
• A checked-in calibration baseline and reviewed regression budget exist.
• Repeated local and CI runs demonstrate that the chosen threshold exceeds ordinary noise without masking material regression.
• Tests reject malformed, partial, or semantically incomplete result records.

[coderabbitai]

🔗 Related PRs

#652 - Make unit tests resource-deterministic [merged]
#732 - Fix: anchor state-cache payloads against Git GC [merged]
#751 - feat(materialize): add retained CAS materialization handles [merged]
#754 - Resume exact materializations from retained git-cas roots [merged]
#755 - fix: retain bounded materialization roots [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!