Limit HGBT threads on small datasets#34150

Open

hanhan761 wants to merge 1 commit into

scikit-learn:mainfrom

hanhan761:hgb-small-data-threads

hanhan761 commented May 29, 2026

Summary

Limit OpenMP threads used by HistGradientBoosting* during fitting on small datasets.
Add regression coverage for the thread-count heuristic and its fit-time integration.
Add a whats_new efficiency entry for the ensemble module.

Related issues

Part of [META] scalability problems on hybrid CPUs #34128
Helps address HistGradientBoostingClassifier/Regressor 15x slowdown on small data problems compared to disabled OpenMP threading #30662

Test plan

python -m pytest sklearn\ensemble_hist_gradient_boosting\tests -q


          Limit HGBT threads on small datasets

225bc58

github-actions Bot added the module:ensemble label

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

module:ensemble