Issues

• Extend the gradient boosting libraries comparison to include GPU backend comparisons.

  probabl-ai/scikit-learn-benchmarks#184 · Filter by author fcharras opened on Nov 9, 2026

• In dashboards, add an option to filter-out benchmarks whose worst backend performance is below some set threshold

  probabl-ai/scikit-learn-benchmarks#183 · Filter by author fcharras opened on Nov 9, 2026

• Add the AWS L4 host to the PR comparison workflow

  probabl-ai/scikit-learn-benchmarks#182 · Filter by author fcharras opened on Nov 9, 2026
  

• intel-laptop: the balanced power profile throttles clocks by ~40% after sustained all-core work

  probabl-ai/scikit-learn-benchmarks#145 · Filter by author cakedev0 opened on Oct 13, 2026
  1

• Active-wait can let one OpenMP call's thread team pollute a later call's timing (not HGB/predict-specific)

  probabl-ai/scikit-learn-benchmarks#90 · Filter by author cakedev0 opened on Sep 14, 2026

• Process-to-process run time variability on multi-socket runners (possibly NUMA-related)

  probabl-ai/scikit-learn-benchmarks#80 · Filter by author cakedev0 opened on Sep 9, 2026
  3

• Add sparse datasets coming from pre-processed text

  probabl-ai/scikit-learn-benchmarks#4 · Filter by author cakedev0 opened on Jul 31, 2026