This article provides a comprehensive study of different ways to make speed benchmarks of gradient boosted decision trees algorithm. We show main problems of several straight forward ways to make benchmarks, explain, why a speed benchmarking is a challenging task and provide a set of reasonable requirements for a benchmark to be fair and useful.
Cite
@article{arxiv.1810.10380,
title = {Why every GBDT speed benchmark is wrong},
author = {Anna Veronika Dorogush and Vasily Ershov and Dmitriy Kruchinin},
journal= {arXiv preprint arXiv:1810.10380},
year = {2018}
}