The RAPIDS data science framework is a collection of libraries for running end-to-end data science pipelines completely on the GPU. The interaction is designed to have a familiar look and feel to working in Python, but utilizes optimized NVIDIA® CUDA® primitives and high-bandwidth GPU memory under the hood. Below are some links to help getting started with the RAPIDS libraries.
Contains a configurator tool to help you choose between the various methods for installing RAPIDS.
Handy PDF reference guide for handling GPU Data Frames (GDF) with cuDF.
10 Minutes to cuDF
Modeled after 10 Minutes to Pandas, this is a short introduction to cuDF that is geared mainly for new users.
10 Minutes to Dask-XGBoost
A short introduction to XGBoost with a distributed CUDA DataFrame via Dask-cuDF.
Using RAPIDS + XGBoost4J-Spark
Getting started guide and source code for XGBoost4J-Spark, along with notebooks and examples.
Multi-GPU with Dask-cuDF
Overview of using Dask for Multi-GPU cuDF solutions, on both a single machine or multiple GPUs across many machines in a cluster.
Our Collection of Example NoteBooks
A Github repository with our introductory examples of XGBoost, cuML demos, cuGraph demos, and more.
Our Extended Collection of Example NoteBooks
A Github repository with examples of XGBoost, cuML demos, cuGraph demos, and more.