tests: add phase3 measurement and comparison scripts

[jlinenkohl]

PR C Draft: Phase 3 Measurement Tooling

Summary

This PR adds lightweight scripts for repeatable baseline/candidate measurement collection and comparison.

What Changed

Files in this PR:

• tests/phase3/run_phase3_measurements.sh
• tests/phase3/compare_phase3_results.sh
• tests/phase3/.gitignore

Scope

• Tooling only.
• No loader behavior changes.
• No relocation policy changes.

Why

Even when some workloads can be rebuilt with different flags, we still need repeatable evidence for the binaries and runtime paths we actually execute. These scripts provide a reproducible path for collecting and comparing relocation-related timing artifacts before making performance claims.

Validation

• Run baseline capture:
  • bash tests/phase3/run_phase3_measurements.sh baseline 10
• Run candidate capture:
  • bash tests/phase3/run_phase3_measurements.sh candidate 10
• Compare outputs:
  • bash tests/phase3/compare_phase3_results.sh <baseline_csv> <candidate_csv>

Reviewer Ask

Please review as measurement infrastructure to support evidence collection, not as behavioral/runtime policy change.