Inventaire residuel machine-dep timings (issue #10158)
Date : scan check_machine_dep_timing.py --all sur l’ensemble des notebooks du dépôt. Detector summary : {“wallclock”: 220, “distribution_param”: 58, “domain_quantity”: 30, “ambiguous”: 0, “total”: 308}
Classification par categorie
| RUNTIME_MEASURED |
42 |
Valeur runtime drainable – a remplacer par ordre de grandeur ou borne superieure. |
| RUNTIME_HINT |
27 |
Timeout/delay/rate-limit – defensable mais a contextualiser par un commentaire. |
| CONFIG_PARAMETRIC |
118 |
Duree de sample/taille d’evenement – frozen, ne pas toucher. |
| AMBIGUOUS |
121 |
Contexte insuffisant – revue manuelle requise. |
Top 15 familles par drainage potentiel
GenAI/Audio |
24 |
38 |
62 |
Sudoku/Sudoku-13-SymbolicAutomata-CSharp.ipynb |
23 |
5 |
28 |
Probas/Infer |
20 |
26 |
46 |
Search/Applications |
13 |
5 |
18 |
QuantConnect/Python |
12 |
1 |
13 |
Sudoku/Sudoku-03-Genetic-Python.ipynb |
9 |
4 |
13 |
SymbolicAI/Lean |
7 |
2 |
9 |
SymbolicAI/Planners |
6 |
3 |
9 |
Search/Part2-CSP |
5 |
4 |
9 |
Sudoku/Sudoku-18-Comparison-CSharp.ipynb |
5 |
0 |
5 |
Sudoku/Sudoku-18-Comparison-Python.ipynb |
5 |
0 |
5 |
GenAI/Image |
4 |
0 |
4 |
GenAI/Plateformes-Conversationnelles |
4 |
0 |
4 |
GenAI/Texte |
4 |
1 |
5 |
GenAI/Video |
4 |
4 |
8 |
Top des notebooks par drainage reel (RUNTIME_MEASURED)
Sudoku-13-SymbolicAutomata-CSharp.ipynb |
0 |
5 |
18 |
5 |
Infer-2-Gaussian-Mixtures.ipynb |
0 |
0 |
19 |
26 |
Sudoku-03-Genetic-Python.ipynb |
3 |
0 |
6 |
4 |
04-1-Educational-Audio-Content.ipynb |
7 |
0 |
0 |
1 |
04-6-Audiobook-Pipeline.ipynb |
6 |
0 |
0 |
2 |
Sudoku-18-Comparison-CSharp.ipynb |
2 |
0 |
3 |
0 |
Sudoku-18-Comparison-Python.ipynb |
1 |
2 |
2 |
0 |
Infer-1b-Premiers-Modeles.ipynb |
0 |
0 |
4 |
10 |
QC-Py-26-LLM-Trading-Signals.ipynb |
0 |
0 |
4 |
0 |
Sudoku-10-ORTools-CSharp.ipynb |
0 |
0 |
4 |
0 |
App-2b-GraphColoring-CSharp.ipynb |
0 |
0 |
3 |
0 |
Sudoku-18b-Statistical-Comparison-Python.ipynb |
0 |
0 |
3 |
0 |
Lean-07b-Examples-Python.ipynb |
3 |
0 |
0 |
0 |
Planners-8-Temporal.ipynb |
0 |
0 |
3 |
0 |
00-3-API-Endpoints-Configuration.ipynb |
0 |
2 |
0 |
0 |
03-1-Multi-Model-Audio-Comparison.ipynb |
0 |
0 |
2 |
1 |
04-3-Music-Composition-Workflow.ipynb |
1 |
1 |
0 |
4 |
04-7-TTS-Voice-Benchmark.ipynb |
2 |
0 |
0 |
0 |
03-3-Performance-Optimization.ipynb |
0 |
0 |
2 |
0 |
03-Chat-Streaming-QA-OWUI.ipynb |
0 |
2 |
0 |
0 |
Synthese executoire
- Drainable total : 190 findings (RUNTIME_MEASURED + RUNTIME_HINT + AMBIGUOUS)
- Frozen (a ne pas toucher) : 118 findings (CONFIG_PARAMETRIC)
- Ratio drainage : 61.7%
Prochaine tranche : prendre la famille avec le plus de RUNTIME_MEASURED strict (cf. table ci-dessus) et drainer manuellement, puis re-lancer ce script.
Retour au sommet