Add criterion and Python benchmark baselines for doeff-vm

[proboscis]

Summary

• add a reusable Python benchmark workload module plus a rewritten benchmarks/benchmark_runner.py that validates results and writes JSON/CSV artifacts
• add a criterion benchmark for packages/doeff-vm covering module_run and pyvm_fresh baseline cases
• add make bench-python / make bench-vm, a benchmark README, benchmark artifact ignores, and smoke tests for the Python benchmark runner

Verification

• uv run pytest tests/test_benchmark_runner.py packages/doeff-vm/tests/test_package_exports.py packages/doeff-vm/tests/test_pyvm.py::test_module_level_run_basic
  • 6 passed in 0.20s
• make bench-python BENCH_PYTHON_ARGS='--runs 3 --iterations 4 --output-dir benchmarks/results/smoke'
  • public_run:pure: runs=3 iterations=4 | min=0.16ms median=0.17ms mean=0.17ms max=0.19ms
  • public_run:state: runs=3 iterations=4 | min=0.63ms median=0.64ms mean=0.65ms max=0.66ms
  • public_run:state_writer: runs=3 iterations=4 | min=0.96ms median=1.00ms mean=0.99ms max=1.01ms
  • generated artifacts:
    • benchmarks/results/smoke/doeff_vm_benchmark_results.json
    • benchmarks/results/smoke/doeff_vm_benchmark_results.csv
• make bench-vm BENCH_VM_ARGS='--sample-size 10 --measurement-time 0.1 --warm-up-time 0.1 --noplot'
  • doeff_vm_baseline/module_run/pure time: [31.529 µs 34.816 µs 37.774 µs]
  • doeff_vm_baseline/pyvm_fresh/pure time: [30.211 µs 31.943 µs 34.707 µs]
  • doeff_vm_baseline/module_run/state time: [2.6103 ms 2.6416 ms 2.6787 ms]
  • doeff_vm_baseline/pyvm_fresh/state time: [2.5980 ms 2.6341 ms 2.6877 ms]
  • doeff_vm_baseline/module_run/state_writer time: [4.2105 ms 4.6090 ms 5.1725 ms]
  • doeff_vm_baseline/pyvm_fresh/state_writer time: [4.1763 ms 4.3576 ms 4.6610 ms]
  • criterion artifacts confirmed under packages/doeff-vm/target/criterion/doeff_vm_baseline/...
• uv run ruff check benchmarks tests/test_benchmark_runner.py
  • All checks passed!

Issue

• refs vm-perf-bench