sklbench hardware comparison dashboard

1182 benchmark records from 2026-07-02

vanilla sklearn

Python 3.12.13

Packages

  • scikit-learn 1.9.0 (conda)
  • numpy 2.4.6 (conda)
  • scipy 1.17.1 (conda)
  • pandas 2.3.3 (conda)

Threadpools

  • libopenblas 0.3.33 (16 threads)
  • libgomp (16 threads)

Full environment

view pixi env JSON

Hardware

CPU

Intel(R) Core(TM) Ultra X7 358H

X86_64, 16 physical cores, 16 logical CPUs

31 GB RAM

GPU(s)

  • level_zero:gpu:0: Intel(R) Arc(TM) B390 GPU (29 GB)

sklearn-AMD8

Python 3.12.13

Packages

  • scikit-learn 1.9.0 (conda)
  • numpy 2.4.6 (conda)
  • scipy 1.17.1 (conda)
  • pandas 2.3.3 (conda)

Threadpools

  • libopenblas 0.3.33 (8 threads)
  • libgomp (8 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 7413 24-Core Processor

X86_64, 8 physical cores, 8 logical CPUs

47 GB RAM

GPU(s)

  • cuda:0: NVIDIA L4 (22 GB)

sklearnex-cpu-AMD8

Python 3.12.13

Packages

  • scikit-learn 1.8.0 (conda)
  • scikit_learn_intelex 2026.1.0 (pip)
  • daal 2026.1.0 (pip)

Threadpools

  • libopenblas 0.3.33 (8 threads)
  • libgomp (8 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 7413 24-Core Processor

X86_64, 8 physical cores, 8 logical CPUs

47 GB RAM

GPU(s)

  • cuda:0: NVIDIA L4 (22 GB)

sklearn-AMD48

Python 3.12.13

Packages

  • scikit-learn 1.9.0 (conda)
  • numpy 2.4.6 (conda)
  • scipy 1.17.1 (conda)
  • pandas 2.3.3 (conda)

Threadpools

  • libopenblas 0.3.33 (48 threads)
  • libgomp (48 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 9555P 64-Core Processor

X86_64, 48 physical cores, 48 logical CPUs

94 GB RAM

GPU(s)

No GPU detected.

sklearnex-cpu-AMD48

Python 3.12.13

Packages

  • scikit-learn 1.8.0 (conda)
  • scikit_learn_intelex 2026.1.0 (pip)
  • daal 2026.1.0 (pip)

Threadpools

  • libopenblas 0.3.33 (48 threads)
  • libgomp (48 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 9555P 64-Core Processor

X86_64, 48 physical cores, 48 logical CPUs

94 GB RAM

GPU(s)

No GPU detected.

sklearn-cupy-cuda-L4

Python 3.12.13

Array API active

Packages

  • scikit-learn 1.9.0 (conda)
  • cupy 14.1.1 (conda)

Threadpools

  • libopenblas 0.3.33 (8 threads)
  • libgomp (8 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 7413 24-Core Processor

X86_64, 8 physical cores, 8 logical CPUs

47 GB RAM

GPU(s)

  • cuda:0: NVIDIA L4 (22 GB)

sklearn-torch-cuda-L4

Python 3.12.13

Array API active

Packages

  • scikit-learn 1.9.0 (conda)
  • torch 2.10.0+cu129 (pip)

Threadpools

  • libopenblas 0.3.33 (8 threads)
  • libgomp (8 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 7413 24-Core Processor

X86_64, 8 physical cores, 8 logical CPUs

47 GB RAM

GPU(s)

  • cuda:0: NVIDIA L4 (22 GB)

linear / fit (128 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🔁 Number of iteration differs: this might mean algorithms differ - LogisticRegression (8)

linear / predict (128 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🔁 Number of iteration differs: this might mean algorithms differ - LogisticRegression (8)
Detailed results

tree-based / fit (360 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🧺 Scikit-learn intelex uses binning & histogram-based splits while scikit-learn doesn't - ExtraTreesClassifier (30), ExtraTreesRegressor (30), RandomForestClassifier (30), RandomForestRegressor (30)
◇ Metrics differ from the baseline (119 points)

tree-based / predict (360 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🧺 Scikit-learn intelex uses binning & histogram-based splits while scikit-learn doesn't - ExtraTreesClassifier (30), ExtraTreesRegressor (30), RandomForestClassifier (30), RandomForestRegressor (30)
Detailed results

clustering / fit (12 points)

● Metrics and benchmark setup match the baseline

clustering / predict (12 points)

● Metrics and benchmark setup match the baseline
Detailed results

196 benchmark records from 2026-07-02

vanilla sklearn

Python 3.12.13

Packages

  • scikit-learn 1.9.0 (conda)
  • numpy 2.4.6 (conda)
  • scipy 1.17.1 (conda)
  • pandas 2.3.3 (conda)

Threadpools

  • libopenblas 0.3.33 (16 threads)
  • libgomp (16 threads)

Full environment

view pixi env JSON

Hardware

CPU

Intel(R) Core(TM) Ultra X7 358H

X86_64, 16 physical cores, 16 logical CPUs

31 GB RAM

GPU(s)

  • level_zero:gpu:0: Intel(R) Arc(TM) B390 GPU (29 GB)

sklearnex-gpu-B390

Python 3.12.13

Packages

  • scikit-learn 1.8.0 (conda)
  • scikit_learn_intelex 2026.1.0 (pip)
  • scikit_learn_intelex_gpu 2026.1.0 (pip)
  • dpnp 0.20.0 (pip)
  • daal 2026.1.0 (pip)

Threadpools

  • libopenblas 0.3.33 (16 threads)
  • libgomp (16 threads)

Full environment

view pixi env JSON

Hardware

CPU

Intel(R) Core(TM) Ultra X7 358H

X86_64, 16 physical cores, 16 logical CPUs

31 GB RAM

GPU(s)

  • level_zero:gpu:0: Intel(R) Arc(TM) B390 GPU (29 GB)

sklearn-dpnp-gpu-B390

Python 3.12.13

Array API active

Packages

  • scikit-learn 1.9.0 (conda)
  • dpnp 0.19.1 (pip)

Threadpools

  • libopenblas 0.3.33 (16 threads)
  • libgomp (16 threads)

Full environment

view pixi env JSON

Hardware

CPU

Intel(R) Core(TM) Ultra X7 358H

X86_64, 16 physical cores, 16 logical CPUs

31 GB RAM

GPU(s)

  • level_zero:gpu:0: Intel(R) Arc(TM) B390 GPU (29 GB)

sklearn-torch-xpu-B390

Python 3.12.13

Array API active

Packages

  • scikit-learn 1.9.0 (conda)
  • torch 2.12.1+xpu (pip)

Threadpools

  • libopenblas 0.3.33 (16 threads)
  • libgomp (16 threads)

Full environment

view pixi env JSON

Hardware

CPU

Intel(R) Core(TM) Ultra X7 358H

X86_64, 16 physical cores, 16 logical CPUs

31 GB RAM

GPU(s)

  • level_zero:gpu:0: Intel(R) Arc(TM) B390 GPU (29 GB)

sklearn-cupy-cuda-L4

Python 3.12.13

Array API active

Packages

  • scikit-learn 1.9.0 (conda)
  • cupy 14.1.1 (conda)

Threadpools

  • libopenblas 0.3.33 (8 threads)
  • libgomp (8 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 7413 24-Core Processor

X86_64, 8 physical cores, 8 logical CPUs

47 GB RAM

GPU(s)

  • cuda:0: NVIDIA L4 (22 GB)

sklearn-torch-cuda-L4

Python 3.12.13

Array API active

Packages

  • scikit-learn 1.9.0 (conda)
  • torch 2.10.0+cu129 (pip)

Threadpools

  • libopenblas 0.3.33 (8 threads)
  • libgomp (8 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 7413 24-Core Processor

X86_64, 8 physical cores, 8 logical CPUs

47 GB RAM

GPU(s)

  • cuda:0: NVIDIA L4 (22 GB)

fit / speed-up vs vanilla sklearn (70 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🔁 Number iteration of iteration not reported for this variant - LogisticRegression (4)
  • ↩ Some operations fell back to CPU according to benchmark logs - Ridge (4), RidgeClassifier (4)
  • 🔁 Number of iteration differs: this might mean algorithms differ - LogisticRegression (4)

predict / speed-up vs vanilla sklearn (70 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🔁 Number iteration of iteration not reported for this variant - LogisticRegression (4)
  • ↩ Some operations fell back to CPU according to benchmark logs - Ridge (4), RidgeClassifier (4)
  • 🔁 Number of iteration differs: this might mean algorithms differ - LogisticRegression (4)
Detailed results

1428 benchmark records from 2026-07-02

vanilla sklearn

Python 3.12.13

Packages

  • scikit-learn 1.9.0 (conda)
  • numpy 2.4.6 (conda)
  • scipy 1.17.1 (conda)
  • pandas 2.3.3 (conda)

Threadpools

  • libopenblas 0.3.33 (16 threads)
  • libgomp (16 threads)

Full environment

view pixi env JSON

Hardware

CPU

Intel(R) Core(TM) Ultra X7 358H

X86_64, 16 physical cores, 16 logical CPUs

31 GB RAM

GPU(s)

  • level_zero:gpu:0: Intel(R) Arc(TM) B390 GPU (29 GB)

sklearnex-cpu-laptop

Python 3.12.13

Packages

  • scikit-learn 1.8.0 (conda)
  • scikit_learn_intelex 2026.1.0 (pip)
  • daal 2026.1.0 (pip)

Threadpools

  • libopenblas 0.3.33 (16 threads)
  • libgomp (16 threads)

Full environment

view pixi env JSON

Hardware

CPU

Intel(R) Core(TM) Ultra X7 358H

X86_64, 16 physical cores, 16 logical CPUs

31 GB RAM

GPU(s)

  • level_zero:gpu:0: Intel(R) Arc(TM) B390 GPU (29 GB)

sklearn-AMD8

Python 3.12.13

Packages

  • scikit-learn 1.9.0 (conda)
  • numpy 2.4.6 (conda)
  • scipy 1.17.1 (conda)
  • pandas 2.3.3 (conda)

Threadpools

  • libopenblas 0.3.33 (8 threads)
  • libgomp (8 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 7413 24-Core Processor

X86_64, 8 physical cores, 8 logical CPUs

47 GB RAM

GPU(s)

  • cuda:0: NVIDIA L4 (22 GB)

sklearnex-cpu-AMD8

Python 3.12.13

Packages

  • scikit-learn 1.8.0 (conda)
  • scikit_learn_intelex 2026.1.0 (pip)
  • daal 2026.1.0 (pip)

Threadpools

  • libopenblas 0.3.33 (8 threads)
  • libgomp (8 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 7413 24-Core Processor

X86_64, 8 physical cores, 8 logical CPUs

47 GB RAM

GPU(s)

  • cuda:0: NVIDIA L4 (22 GB)

sklearn-AMD48

Python 3.12.13

Packages

  • scikit-learn 1.9.0 (conda)
  • numpy 2.4.6 (conda)
  • scipy 1.17.1 (conda)
  • pandas 2.3.3 (conda)

Threadpools

  • libopenblas 0.3.33 (48 threads)
  • libgomp (48 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 9555P 64-Core Processor

X86_64, 48 physical cores, 48 logical CPUs

94 GB RAM

GPU(s)

No GPU detected.

sklearnex-cpu-AMD48

Python 3.12.13

Packages

  • scikit-learn 1.8.0 (conda)
  • scikit_learn_intelex 2026.1.0 (pip)
  • daal 2026.1.0 (pip)

Threadpools

  • libopenblas 0.3.33 (48 threads)
  • libgomp (48 threads)

Full environment

view pixi env JSON

Hardware

CPU

AMD EPYC 9555P 64-Core Processor

X86_64, 48 physical cores, 48 logical CPUs

94 GB RAM

GPU(s)

No GPU detected.

linear / fit (128 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🔁 Number of iteration differs: this might mean algorithms differ - LogisticRegression (14)

linear / predict (128 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🔁 Number of iteration differs: this might mean algorithms differ - LogisticRegression (14)
Detailed results

tree-based / fit (480 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🧺 Scikit-learn intelex uses binning & histogram-based splits while scikit-learn doesn't - ExtraTreesClassifier (45), ExtraTreesRegressor (45), RandomForestClassifier (45), RandomForestRegressor (45)
◇ Metrics differ from the baseline (165 points)

tree-based / predict (480 points)

● Metrics and benchmark setup match the baseline
■ Metrics match the baseline, but some comparison details are worth reporting:
  • 🧺 Scikit-learn intelex uses binning & histogram-based splits while scikit-learn doesn't - ExtraTreesClassifier (45), ExtraTreesRegressor (45), RandomForestClassifier (45), RandomForestRegressor (45)
Detailed results

clustering / fit (15 points)

● Metrics and benchmark setup match the baseline

clustering / predict (15 points)

● Metrics and benchmark setup match the baseline
Detailed results