Learning Random Forests on the GPU
Yi-Sheng Liao, Alex Rubinsteyn, Russell Power, Jinyang Li · 2013
Random Forests are a popular and powerful machine learning technique, with several fast multi-core CPU implementations. Since many other machine learn-ing methods have seen impressive speedups from GPU implementations, applying GPU acceleration to random forests seems like a natural fit. Previous attempts to use GPUs have relied on coarse-grained task parallelism and have yielded incon-clusive or unsatisfying results. We introduce CudaTree, a GPU Random Forest implementation which adaptively switches between data and task parallelism. We show that, for larger datasets, this algorithm is faster than highly tuned multi-core CPU implementations. 1