#9 Add paired A/B suite comparison

[iray-tno]

Summary

• add prepkit test suite compare A B cases for paired numeric comparisons
• support --runs, --workers, --score-mode, preprocessing, and Rust/C++ compilation paths
• report per-pair diff, A/B/tie sign consistency, mean diff, and diff stdev

Tests

• uv run pytest tests/test_test_command.py -q
• uv run pytest -q