# AMD Ryzen 9 7950X3D (Zen 4), Debian 13, 4 CPUs (taskset -c 0-3), Python 3.12.15, scikit-learn 1.9.1, numpy 2.5.4, pandas 3.0.6, joblib 1.6.0, 12 October 2026
# official wheel: lightgbm 4.7.0 from PyPI

$ python pi_cost.py count work
predict calls: 251 (1 + 50 features x 5 repeats), 10000 rows each, 300 trees: 753 million tree walks
one predict_proba over 10000 rows, 1 thread(s): 179.4 ms (median of 7)
one predict_proba over 10000 rows, 4 thread(s): 46.4 ms (median of 7)

$ python pi_cost.py all work 3
N = 4 CPUs, 3 rounds per setting, median wall time of the permutation_importance call
pandas, n_jobs=1, 1 thread                          46.01 s   5f242b6b5bd26ee0 same result
numpy, n_jobs=1, 1 thread                           45.87 s   5f242b6b5bd26ee0 same result
numpy, n_jobs=1, N threads                          12.45 s   5f242b6b5bd26ee0 same result
numpy, n_jobs=N, 1 thread                           13.80 s   5f242b6b5bd26ee0 same result
pandas, n_jobs=N, 1 thread                          13.79 s   5f242b6b5bd26ee0 same result
numpy, n_jobs=N, LightGBM default threads           15.18 s   5f242b6b5bd26ee0 same result
numpy, n_jobs=N, 1 thread, max_samples=0.5           7.96 s   b756cc6b72a439bd DIFFERENT result
numpy, n_jobs=N, 1 thread, n_repeats=2               6.75 s   df04964ce9cc6f9e DIFFERENT result
numpy float32, n_jobs=N, 1 thread                   13.89 s   5f242b6b5bd26ee0 same result

$ python pi_cost.py all work 3 "pandas, n_jobs=1, LightGBM default threads" "numpy, n_jobs=2, N/2 threads"
N = 4 CPUs, 3 rounds per setting, median wall time of the permutation_importance call
pandas, n_jobs=1, LightGBM default threads          12.97 s   5f242b6b5bd26ee0 same result
numpy, n_jobs=2, N/2 threads                        14.47 s   5f242b6b5bd26ee0 same result