feat: allow targeted strategy smoke checks

[kaozyn]

PR kind

• fix
• feat

Summary

• add optional transform-name arguments to python -m strategy.smoke
• allow contributors to run targeted smoke checks like python -m strategy.smoke rsvd
• add CPU-safe tests for transform selection and unknown-name handling
• use this PR to exercise the new feat lane automation end to end

Validation

• uv run --extra test python -m pytest tests/test_smoke_cli.py -q
• uv run --extra test python -m py_compile $(find matmul strategy eval tests examples -name '*.py')
• uv run --extra test python -m pytest tests/ strategy/tests/ eval/tests/ -q

GPU Result (required for feat PRs only)

metric	value
accuracy	0.0022
time complexity	O(N^2·M), with default M=N//8 behaving near O(N^3)
latency	1093.16 ms
VRAM usage	713.08 MiB

Regime measured: N=12000, dtype=fp32, fill=full-rank, rank M=1500, device=RTX 4090

Raw scorecard (paste python -m eval … output or --json)

RESULT_JSON: {"config":{"n":12000,"pairs":3,"dtype":"fp32","rank_m":1500,"fill":"random","accuracy_floor":0.8,"vram_unit":"gib","device":"NVIDIA GeForce RTX 4090 (PyTorch/CUDA)","seed":11},"complexity":{"normal":"O(N^3)","smart":"O(N^2 * M)  (M=min(N,max(64,N/8))=1500 -> ~O(N^3) when M grows with N)"},"exact":{"latency_s":0.6397263603284955,"peak_vram_mib":1658.125},"transforms":{"rsvd":{"accuracy":0.0021774302182113368,"latency_s":1.0931598281798263,"peak_vram_mib":713.07958984375,"flop_ratio_vs_exact":1.7716262975778547,"gated":true,"improvement":false,"score":0.0}},"ranking":["rsvd"],"best":"rsvd"}

Checklist

• CPU-safe validation passed (strategy.smoke if relevant, plus pytest tests/ strategy/tests/ eval/tests/).
• If this is a feat PR, I ran the scorer on unseen couples — no hardcoding of seeds/matrices.
• If this is a feat PR, accuracy and latency come from the same run at the same dtype.
• If this is a feat PR, this is an improvement (every cost axis down, accuracy held) or I state honestly which axis it trades — see the one rule in CONTRIBUTING.md.
• Correctness gates pass: python eval/tests/test_eval.py, python strategy/tests/test_subspace.py, python tests/test_correctness.py.
• If this is a feat PR, I named the device and dtype so a reviewer can reproduce the numbers.

[github-actions]

Gate chain passed. This PR is queued for the next batched GPU evaluation window.

[github-actions]

GPU evaluation complete on RTX 5090 (mock).

• Verdict: eval:none
  Reason: gain -1.9% does not clear the 2% significance floor
• Track: full-rank / transform mock_transform
• Accuracy: 0.94
• Latency: 0.021 seconds
• Peak VRAM: 1536.0 MiB
• FLOP ratio vs exact: 2.5
• Score: 10.2
• Seed: 8878683