from sklearn_benchmarks.reporting.hp_match import HpMatchReporting
from sklearn_benchmarks.utils import default_run_dir, default_report_config
from pathlib import Path
import pandas as pd
pd.set_option('display.max_colwidth', None)
pd.set_option('display.max_columns', None)
pd.set_option('display.max_rows', None)

scikit-learn-intelex (Intel® oneAPI) vs. scikit-learn

run_dir = default_run_dir()
report_config = default_report_config("scikit_learn_intelex_vs_scikit_learn")

# Parameters
report_config = {
    "parameters": {"n_cols": 2},
    "estimators": [
        {
            "title": "KNeighborsClassifier (brute force)",
            "name": "KNeighborsClassifier_brute_force",
            "split_bars_by_parameters": ["n_jobs"],
        },
        {
            "title": "KNeighborsClassifier (KD tree)",
            "name": "KNeighborsClassifier_kd_tree",
            "split_bars_by_parameters": ["n_jobs"],
        },
        {"title": "KMeans (tall)", "name": "KMeans_tall"},
        {"title": "KMeans (short)", "name": "KMeans_short"},
        {"title": "Logistic Regression", "name": "LogisticRegression"},
        {"title": "LinearRegression", "name": "LinearRegression"},
        {"title": "Ridge", "name": "Ridge"},
        {"title": "TSNE", "name": "TSNE"},
        {"title": "PCA", "name": "PCA"},
    ],
    "log_scale": True,
    "other_library": "sklearnex",
    "benchmarking_estimator_parameters": {
        "sklearn_KNeighborsClassifier_brute_force": [
            "n_neighbors",
            "algorithm",
            "n_jobs",
        ],
        "sklearnex_KNeighborsClassifier_brute_force": [],
        "sklearn_KNeighborsClassifier_kd_tree": ["n_neighbors", "algorithm", "n_jobs"],
        "sklearnex_KNeighborsClassifier_kd_tree": [],
        "sklearn_KMeans_tall": [
            "init",
            "algorithm",
            "n_clusters",
            "max_iter",
            "n_init",
            "tol",
        ],
        "sklearnex_KMeans_tall": [],
        "sklearn_KMeans_short": [
            "init",
            "algorithm",
            "n_clusters",
            "max_iter",
            "n_init",
            "tol",
        ],
        "sklearnex_KMeans_short": [],
        "sklearn_LogisticRegression": [
            "penalty",
            "dual",
            "tol",
            "C",
            "fit_intercept",
            "intercept_scaling",
            "class_weight",
            "random_state",
            "solver",
            "max_iter",
            "multi_class",
            "verbose",
            "warm_start",
            "n_jobs",
            "l1_ratio",
        ],
        "sklearnex_LogisticRegression": [],
        "sklearn_LinearRegression": [],
        "sklearnex_LinearRegression": [],
        "sklearn_Ridge": ["alpha"],
        "sklearnex_Ridge": [],
        "sklearn_TSNE": [],
        "sklearnex_TSNE": [],
        "sklearn_PCA": ["n_components"],
        "sklearnex_PCA": [],
        "sklearn_HistGradientBoostingClassifier_best": [
            "learning_rate",
            "n_iter_no_change",
            "max_leaf_nodes",
            "max_bins",
            "min_samples_leaf",
            "max_iter",
        ],
        "sklearn_HistGradientBoostingClassifier": [
            "learning_rate",
            "n_iter_no_change",
            "max_leaf_nodes",
            "max_bins",
            "min_samples_leaf",
            "max_iter",
            "early_stopping",
        ],
        "lightgbm": [
            "learning_rate",
            "num_leaves",
            "max_bin",
            "min_data_in_leaf",
            "bagging_fraction",
            "feature_fraction_bynode",
            "n_estimators",
        ],
        "xgboost": [
            "validate_parameters",
            "tree_method",
            "use_label_encoder",
            "learning_rate",
            "max_leaves",
            "max_bin",
            "min_child_weight",
            "subsample",
            "colsample_bynode",
            "n_estimators",
        ],
        "catboost_lossguide": [
            "allow_writing_files",
            "grow_policy",
            "bootstrap_type",
            "learning_rate",
            "max_leaves",
            "border_count",
            "min_data_in_leaf",
            "subsample",
            "rsm",
            "max_depth",
            "iterations",
        ],
        "catboost_symmetric": [
            "allow_writing_files",
            "grow_policy",
            "bootstrap_type",
            "learning_rate",
            "border_count",
            "min_data_in_leaf",
            "subsample",
            "rsm",
            "max_depth",
            "iterations",
        ],
    },
}
run_dir = "./results/local/20220322T182254/"

run_dir = Path(run_dir)

reporting = HpMatchReporting(**report_config, run_dir=run_dir)
reporting.make_report()

We assume here there is a perfect match between the hyperparameters of both librairies. For a given set of parameters and a given dataset, we compute the speed-up time scikit-learn / time sklearnex. For instance, a speed-up of 2 means that sklearnex is twice as fast as scikit-learn for a given set of parameters and a given dataset.

KNeighborsClassifier (brute force) ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters: algorithm=brute.

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_jobs	n_neighbors	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	100000	100000	100	1.072	0.018	0.075	0.0	-1	1	0.047	0.001	22.613	22.615
3	Download	Download	fit	100000	100000	100	1.009	0.011	0.079	0.0	-1	5	0.044	0.000	23.121	23.121
6	Download	Download	fit	100000	100000	100	1.023	0.016	0.078	0.0	1	100	0.044	0.000	23.202	23.203
9	Download	Download	fit	100000	100000	100	1.024	0.018	0.078	0.0	-1	100	0.047	0.000	21.703	21.703
12	Download	Download	fit	100000	100000	100	1.025	0.018	0.078	0.0	1	5	0.047	0.000	21.734	21.734
15	Download	Download	fit	100000	100000	100	1.018	0.017	0.079	0.0	1	1	0.045	0.002	22.514	22.532
18	Download	Download	fit	100000	100000	2	0.041	0.001	0.039	0.0	-1	1	0.008	0.000	5.386	5.390
21	Download	Download	fit	100000	100000	2	0.041	0.000	0.039	0.0	-1	5	0.008	0.000	5.408	5.410
24	Download	Download	fit	100000	100000	2	0.040	0.001	0.040	0.0	1	100	0.008	0.000	5.290	5.292
27	Download	Download	fit	100000	100000	2	0.041	0.001	0.039	0.0	-1	100	0.008	0.000	5.431	5.433
30	Download	Download	fit	100000	100000	2	0.045	0.000	0.035	0.0	1	5	0.008	0.000	5.980	5.983
33	Download	Download	fit	100000	100000	2	0.045	0.001	0.036	0.0	1	1	0.008	0.000	5.898	5.899

predict

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_jobs	n_neighbors	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
1	Download	Download	predict	100000	1000	100	2.209	0.076	0.000	0.002	-1	1	0.163	0.001	13.527	13.527
2	Download	Download	predict	100000	1	100	0.021	0.002	0.000	0.021	-1	1	0.008	0.000	2.709	2.712
4	Download	Download	predict	100000	1000	100	2.693	0.081	0.000	0.003	-1	5	0.159	0.001	16.925	16.925
5	Download	Download	predict	100000	1	100	0.022	0.003	0.000	0.022	-1	5	0.008	0.000	2.830	2.831
7	Download	Download	predict	100000	1000	100	1.743	0.011	0.000	0.002	1	100	0.195	0.001	8.947	8.947
8	Download	Download	predict	100000	1	100	0.018	0.000	0.000	0.018	1	100	0.008	0.000	2.319	2.321
10	Download	Download	predict	100000	1000	100	2.435	0.042	0.000	0.002	-1	100	0.195	0.002	12.479	12.479
11	Download	Download	predict	100000	1	100	0.022	0.002	0.000	0.022	-1	100	0.008	0.000	2.841	2.842
13	Download	Download	predict	100000	1000	100	1.725	0.003	0.000	0.002	1	5	0.158	0.001	10.903	10.903
14	Download	Download	predict	100000	1	100	0.018	0.000	0.000	0.018	1	5	0.008	0.000	2.389	2.392
16	Download	Download	predict	100000	1000	100	1.115	0.005	0.001	0.001	1	1	0.158	0.001	7.062	7.062
17	Download	Download	predict	100000	1	100	0.018	0.000	0.000	0.018	1	1	0.008	0.001	2.236	2.244
19	Download	Download	predict	100000	1000	2	1.620	0.036	0.000	0.002	-1	1	0.024	0.000	66.663	66.666
20	Download	Download	predict	100000	1	2	0.004	0.002	0.000	0.004	-1	1	0.001	0.000	5.176	5.235
22	Download	Download	predict	100000	1000	2	2.285	0.022	0.000	0.002	-1	5	0.025	0.001	90.171	90.199
23	Download	Download	predict	100000	1	2	0.007	0.002	0.000	0.007	-1	5	0.001	0.000	9.380	9.538
25	Download	Download	predict	100000	1000	2	1.690	0.013	0.000	0.002	1	100	0.057	0.000	29.808	29.808
26	Download	Download	predict	100000	1	2	0.002	0.000	0.000	0.002	1	100	0.001	0.000	2.886	2.914
28	Download	Download	predict	100000	1000	2	2.382	0.065	0.000	0.002	-1	100	0.057	0.000	41.941	41.941
29	Download	Download	predict	100000	1	2	0.005	0.002	0.000	0.005	-1	100	0.001	0.000	6.663	6.723
31	Download	Download	predict	100000	1000	2	1.829	0.006	0.000	0.002	1	5	0.025	0.000	72.623	72.625
32	Download	Download	predict	100000	1	2	0.003	0.000	0.000	0.003	1	5	0.001	0.000	3.431	3.458
34	Download	Download	predict	100000	1000	2	1.062	0.006	0.000	0.001	1	1	0.024	0.000	43.678	43.681
35	Download	Download	predict	100000	1	2	0.002	0.000	0.000	0.002	1	1	0.001	0.000	2.204	2.246

Profiling traces can be visualized using Perfetto UI.

KNeighborsClassifier (KD tree) ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters: algorithm=kd_tree.

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_jobs	n_neighbors	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	1000000	1000000	10	2.579	0.037	0.031	0.0	-1	1	0.638	0.003	4.041	4.041
3	Download	Download	fit	1000000	1000000	10	2.616	0.044	0.031	0.0	-1	5	0.642	0.010	4.075	4.075
6	Download	Download	fit	1000000	1000000	10	2.581	0.026	0.031	0.0	1	100	0.649	0.014	3.975	3.976
9	Download	Download	fit	1000000	1000000	10	2.567	0.023	0.031	0.0	-1	100	0.648	0.010	3.960	3.960
12	Download	Download	fit	1000000	1000000	10	2.641	0.048	0.030	0.0	1	5	0.645	0.012	4.096	4.097
15	Download	Download	fit	1000000	1000000	10	2.639	0.023	0.030	0.0	1	1	0.642	0.007	4.110	4.110
18	Download	Download	fit	1000	1000	2	0.000	0.000	0.032	0.0	-1	1	0.001	0.000	0.563	0.568
21	Download	Download	fit	1000	1000	2	0.000	0.000	0.033	0.0	-1	5	0.001	0.000	0.538	0.543
24	Download	Download	fit	1000	1000	2	0.000	0.000	0.032	0.0	1	100	0.001	0.000	0.566	0.576
27	Download	Download	fit	1000	1000	2	0.000	0.000	0.032	0.0	-1	100	0.001	0.000	0.560	0.564
30	Download	Download	fit	1000	1000	2	0.000	0.000	0.032	0.0	1	5	0.001	0.000	0.542	0.548
33	Download	Download	fit	1000	1000	2	0.000	0.000	0.032	0.0	1	1	0.001	0.000	0.570	0.574

predict

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_jobs	n_neighbors	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
1	Download	Download	predict	1000000	1000	10	0.356	0.004	0.000	0.000	-1	1	0.093	0.001	3.813	3.813
2	Download	Download	predict	1000000	1	10	0.003	0.001	0.000	0.003	-1	1	0.000	0.000	9.272	10.033
4	Download	Download	predict	1000000	1000	10	0.655	0.005	0.000	0.001	-1	5	0.170	0.005	3.854	3.856
5	Download	Download	predict	1000000	1	10	0.003	0.000	0.000	0.003	-1	5	0.000	0.000	8.601	9.191
7	Download	Download	predict	1000000	1000	10	4.109	0.023	0.000	0.004	1	100	0.508	0.017	8.085	8.089
8	Download	Download	predict	1000000	1	10	0.002	0.001	0.000	0.002	1	100	0.001	0.000	4.136	4.300
10	Download	Download	predict	1000000	1000	10	2.220	0.012	0.000	0.002	-1	100	0.503	0.007	4.413	4.413
11	Download	Download	predict	1000000	1	10	0.004	0.001	0.000	0.004	-1	100	0.001	0.000	6.945	7.324
13	Download	Download	predict	1000000	1000	10	1.250	0.013	0.000	0.001	1	5	0.165	0.002	7.602	7.602
14	Download	Download	predict	1000000	1	10	0.001	0.000	0.000	0.001	1	5	0.000	0.000	3.928	4.195
16	Download	Download	predict	1000000	1000	10	0.651	0.010	0.000	0.001	1	1	0.093	0.001	6.991	6.991
17	Download	Download	predict	1000000	1	10	0.001	0.000	0.000	0.001	1	1	0.000	0.000	3.813	4.042
19	Download	Download	predict	1000	1000	2	0.020	0.001	0.001	0.000	-1	1	0.001	0.000	40.359	41.108
20	Download	Download	predict	1000	1	2	0.002	0.000	0.000	0.002	-1	1	0.000	0.000	15.793	18.216
22	Download	Download	predict	1000	1000	2	0.021	0.001	0.001	0.000	-1	5	0.001	0.000	26.981	27.526
23	Download	Download	predict	1000	1	2	0.002	0.000	0.000	0.002	-1	5	0.000	0.000	14.742	17.338
25	Download	Download	predict	1000	1000	2	0.030	0.000	0.001	0.000	1	100	0.004	0.000	7.150	7.169
26	Download	Download	predict	1000	1	2	0.001	0.000	0.000	0.001	1	100	0.000	0.000	3.662	4.254
28	Download	Download	predict	1000	1000	2	0.030	0.000	0.001	0.000	-1	100	0.004	0.000	6.805	6.844
29	Download	Download	predict	1000	1	2	0.002	0.000	0.000	0.002	-1	100	0.000	0.000	14.758	16.816
31	Download	Download	predict	1000	1000	2	0.019	0.000	0.001	0.000	1	5	0.001	0.000	22.601	23.297
32	Download	Download	predict	1000	1	2	0.001	0.000	0.000	0.001	1	5	0.000	0.000	3.760	4.420
34	Download	Download	predict	1000	1000	2	0.017	0.000	0.001	0.000	1	1	0.000	0.000	36.654	38.035
35	Download	Download	predict	1000	1	2	0.001	0.000	0.000	0.001	1	1	0.000	0.000	4.020	4.502

Profiling traces can be visualized using Perfetto UI.

KMeans (tall) ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters: algorithm=full, n_clusters=3, max_iter=30, n_init=1, tol=1e-16.

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_iter	init	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	1000000	1000000	2	0.439	0.002	0.036	0.0	30	random	0.354	0.024	1.239	1.241
3	Download	Download	fit	1000000	1000000	2	0.514	0.003	0.031	0.0	30	k-means++	0.390	0.020	1.316	1.318
6	Download	Download	fit	1000000	1000000	100	4.332	0.197	0.185	0.0	30	random	2.485	0.010	1.743	1.743
9	Download	Download	fit	1000000	1000000	100	4.620	0.058	0.173	0.0	30	k-means++	2.626	0.009	1.759	1.759

predict

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_iter	init	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
1	Download	Download	predict	1000000	1000	2	0.0	0.0	0.070	0.0	30	random	0.0	0.0	1.151	1.288
2	Download	Download	predict	1000000	1	2	0.0	0.0	0.000	0.0	30	random	0.0	0.0	1.306	1.486
4	Download	Download	predict	1000000	1000	2	0.0	0.0	0.075	0.0	30	k-means++	0.0	0.0	1.048	1.168
5	Download	Download	predict	1000000	1	2	0.0	0.0	0.000	0.0	30	k-means++	0.0	0.0	1.148	1.334
7	Download	Download	predict	1000000	1000	100	0.0	0.0	1.842	0.0	30	random	0.0	0.0	1.425	1.515
8	Download	Download	predict	1000000	1	100	0.0	0.0	0.004	0.0	30	random	0.0	0.0	1.261	1.366
10	Download	Download	predict	1000000	1000	100	0.0	0.0	1.894	0.0	30	k-means++	0.0	0.0	1.195	1.255
11	Download	Download	predict	1000000	1	100	0.0	0.0	0.004	0.0	30	k-means++	0.0	0.0	1.153	1.336

Profiling traces can be visualized using Perfetto UI.

KMeans (short) ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters: algorithm=full, n_clusters=300, max_iter=20, n_init=1, tol=1e-16.

Mismatches between validation scores

WARNING! Mismatch between validation scores for 2 predictions (25.0%).
The observed differences can be found in the diff_adjusted_rand_scores column.
The chosen difference threshold is 0.01 for adjusted_rand_score.
The biggest difference is 0.077 for adjusted_rand_score.
See details in the dataframe below.

	estimator	library	diff_adjusted_rand_scores	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_iter	algorithm	init	max_iter	n_clusters	n_init	tol	adjusted_rand_score_sklearn	mean_duration_sklearnex	std_duration_sklearnex	adjusted_rand_score_sklearnex	speedup	std_speedup
7	sklearn_KMeans_short	sklearn	0.074768	predict	10000	1000	100	0.001253	0.000204	0.638494	0.000001	20	full	random	20	300	1	1.000000e-16	0.222701	0.001082	0.000136	0.297469	1.157530	1.166648
10	sklearn_KMeans_short	sklearn	0.076616	predict	10000	1000	100	0.001242	0.000216	0.644047	0.000001	20	full	k-means++	20	300	1	1.000000e-16	0.408030	0.001076	0.000158	0.331414	1.154312	1.166629

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_iter	init	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	10000	10000	2	0.074	0.001	0.002	0.0	20	random	0.029	0.000	2.571	2.572
3	Download	Download	fit	10000	10000	2	0.186	0.002	0.001	0.0	20	k-means++	0.079	0.001	2.348	2.348
6	Download	Download	fit	10000	10000	100	0.181	0.002	0.044	0.0	20	random	0.121	0.002	1.491	1.492
9	Download	Download	fit	10000	10000	100	0.512	0.010	0.016	0.0	20	k-means++	0.309	0.002	1.656	1.656

predict

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_iter	init	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
1	Download	Download	predict	10000	1000	2	0.001	0.0	0.025	0.0	20	random	0.001	0.0	1.021	1.032
2	Download	Download	predict	10000	1	2	0.000	0.0	0.000	0.0	20	random	0.000	0.0	1.357	1.457
4	Download	Download	predict	10000	1000	2	0.001	0.0	0.024	0.0	20	k-means++	0.001	0.0	1.001	1.017
5	Download	Download	predict	10000	1	2	0.000	0.0	0.000	0.0	20	k-means++	0.000	0.0	1.601	1.839
7	Download	Download	predict	10000	1000	100	0.001	0.0	0.638	0.0	20	random	0.001	0.0	1.158	1.167
8	Download	Download	predict	10000	1	100	0.000	0.0	0.003	0.0	20	random	0.000	0.0	1.132	1.206
10	Download	Download	predict	10000	1000	100	0.001	0.0	0.644	0.0	20	k-means++	0.001	0.0	1.154	1.167
11	Download	Download	predict	10000	1	100	0.000	0.0	0.003	0.0	20	k-means++	0.000	0.0	1.211	1.323

Profiling traces can be visualized using Perfetto UI.

Logistic Regression ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters: penalty=l2, dual=False, tol=0.0001, C=1.0, fit_intercept=True, intercept_scaling=1.0, class_weight=nan, random_state=nan, solver=lbfgs, max_iter=100, multi_class=auto, verbose=0, warm_start=False, n_jobs=nan, l1_ratio=nan.

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_iter	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	1000000	1000000	100	9.529	0.066	0.084	0.000	[20]	1.679	0.003	5.676	5.676
3	Download	Download	fit	1000	1000	10000	0.688	0.046	0.116	0.001	[27]	0.693	0.057	0.993	0.996

predict

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_iter	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
1	Download	Download	predict	1000000	1000	100	0.000	0.0	2.893	0.0	[20]	0.000	0.0	0.717	0.765
2	Download	Download	predict	1000000	1	100	0.000	0.0	0.015	0.0	[20]	0.000	0.0	0.248	0.266
4	Download	Download	predict	1000	100	10000	0.002	0.0	5.204	0.0	[27]	0.003	0.0	0.527	0.530
5	Download	Download	predict	1000	1	10000	0.000	0.0	0.962	0.0	[27]	0.001	0.0	0.112	0.115

Profiling traces can be visualized using Perfetto UI.

LinearRegression ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters:

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	copy_X	fit_intercept	normalize	positive	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	1000	1000	10000	1.261	0.012	0.063	0.001	True	True	deprecated	False	1.422	0.015	0.887	0.887
3	Download	Download	fit	1000000	1000000	100	7.059	0.044	0.113	0.000	True	True	deprecated	False	0.174	0.001	40.480	40.481

predict

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	copy_X	fit_intercept	normalize	positive	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
1	Download	Download	predict	1000	1000	10000	0.01	0.0	8.346	0.0	True	True	deprecated	False	0.016	0.0	0.595	0.595
2	Download	Download	predict	1000	10	10000	0.00	0.0	6.137	0.0	True	True	deprecated	False	0.000	0.0	0.469	0.491
4	Download	Download	predict	1000000	1000	100	0.00	0.0	6.170	0.0	True	True	deprecated	False	0.000	0.0	0.310	0.321
5	Download	Download	predict	1000000	10	100	0.00	0.0	0.133	0.0	True	True	deprecated	False	0.000	0.0	0.444	0.473

Profiling traces can be visualized using Perfetto UI.

Ridge ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters: alpha=1e-06.

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	1000	1000	10000	0.159	0.001	0.504	0.0	0.167	0.001	0.952	0.952
3	Download	Download	fit	1000000	1000000	100	1.110	0.009	0.721	0.0	0.212	0.000	5.244	5.244

predict

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
1	Download	Download	predict	1000	1000	10000	0.01	0.0	8.367	0.0	0.016	0.0	0.588	0.588
2	Download	Download	predict	1000	10	10000	0.00	0.0	5.666	0.0	0.000	0.0	0.489	0.521
4	Download	Download	predict	1000000	1000	100	0.00	0.0	4.625	0.0	0.000	0.0	0.575	0.620
5	Download	Download	predict	1000000	10	100	0.00	0.0	0.152	0.0	0.000	0.0	0.329	0.368

Profiling traces can be visualized using Perfetto UI.

TSNE ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters:

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_iter	angle	early_exaggeration	init	learning_rate	method	metric	min_grad_norm	n_components	n_iter_without_progress	perplexity	square_distances	verbose	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	1000	1000	3	3.395	0.168	0.0	0.003	1000	0.5	12.0	warn	warn	barnes_hut	euclidean	0.0	2	300	30.0	legacy	0	3.409	0.236	0.996	0.998

Profiling traces can be visualized using Perfetto UI.

PCA ¶

scikit-learn-intelex (2021.20220207.123530) vs. scikit-learn (1.0.2)

Speedup barplots ¶

All estimators share the following parameters: n_components=10.

Raw results ¶

fit

	sklearn_profiling	sklearnex_profiling	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	mean_duration_sklearnex	std_duration_sklearnex	speedup	std_speedup
0	Download	Download	fit	100000	100000	100	0.425	0.008	0.188	0.0	0.027	0.004	15.888	16.104
1	Download	Download	fit	10000	10000	1000	0.356	0.004	0.224	0.0	0.176	0.011	2.025	2.029

Profiling traces can be visualized using Perfetto UI.

Benchmark environment information

System

python	3.8.12 | packaged by conda-forge | (default, Jan 30 2022, 23:42:07) [GCC 9.4.0]
executable	/usr/share/miniconda/envs/sklbench/bin/python
machine	Linux-5.11.0-1028-azure-x86_64-with-glibc2.10

Dependencies

pip	22.0.4
setuptools	60.10.0
sklearn	1.0.2
numpy	1.22.3
scipy	1.8.0
Cython	None
pandas	1.4.1
matplotlib	3.5.1
joblib	1.1.0
threadpoolctl	3.1.0

Threadpool

	user_api	internal_api	prefix	filepath	version	threading_layer	architecture	num_threads
0	blas	openblas	libopenblas	/usr/share/miniconda/envs/sklbench/lib/libopenblasp-r0.3.18.so	0.3.18	pthreads	SkylakeX	2
1	openmp	openmp	libgomp	/usr/share/miniconda/envs/sklbench/lib/libgomp.so.1.0.0	None	NaN	NaN	2

Cpu_count

cpu_count	2
physical_cpu_count	2