Hardware comparison

Pick a baselineThe reference that speed-ups are computed against. and a comparison machine. The dropdowns only offer CPU vs CPU or GPU vs GPU pairs that have comparable results. Each cell shows the fitTraining a model on data. More in the scikit-learn glossary or predictMaking predictions with a fitted model. More in the scikit-learn glossary speed-up (log scale) per estimator category, one line per build or implementation. The last cell compares the time per fit in hyper-parameter searches, where many fits run in parallel.

On the benchmarked cases:

• The high-end Intel Xeon server mostly helps random forests and extra trees. With scikit-learn, their fit is ~2x to 4x faster than on the Intel Core Ultra laptop. Their predict is close to parity.
• scikit-learn-intelex makes much better use of the server's 172 cores. With it, random forests and extra trees fit ~10x to 15x faster on the server than on the laptop.
• For linear models, the server is often slower than the laptop. A single fit doesn't use many cores well, so the laptop's faster cores probably win.
• Hyper-parameter searches are where the server pays off. Running many candidatesHyperparameter combinations tried by the search. Each one is fitted and scored on every cross-validation split. in parallel fills its cores, and it gets through fits ~10x to 15x faster than the laptop, even for the linear models. That is close to the ratio of physical coresOne hardware core of the CPU. With SMT, it shows up as two logical CPUs. (172 vs 16). The gain is smaller for HistGradientBoosting on small datasets. Random forests and extra trees aren't compared yet, because their forest size differs between the two machines.
• HistGradientBoosting results are hard to interpret, because of known scalability issues on laptops with conda-forgeCommunity-maintained channel of conda packages. Its scikit-learn can be installed with different BLAS and OpenMP libraries. builds and on machines with many cores. See scikit-learn#34764 and the HistGradientBoosting thread-scalability breakdown.
• On the PyPIThe Python Package Index, where pip installs packages from. Its scikit-learn wheels bundle their own OpenMP runtime, and NumPy and SciPy bundle OpenBLAS. build, the Apple M4 CPU is ~1.3x to 2x slower than the Intel Core Ultra laptop.