Welcome to the CUPY-Xarray Notebooks repository for Scipy-2023! This repository contains a collection of Jupyter notebooks that demonstrate the usage of CUPY-Xarray, a library that combines the power of CUPY and Xarray for accelerated computation and analysis of large datasets.

The materials and notebooks in this tutorial is published as a Jupyter book here.

CUPY-Xarray is a Python library that leverages Cupy, a GPU array library, and Xarray, a library for multi-dimensional labeled array computations, to enable fast and efficient data processing on GPUs. By combining the capabilities of Cupy and Xarray, CUPY-Xarray provides a convenient interface for performing accelerated computations and analysis on large multidimensional datasets.

This repository contains the following Jupyter notebooks:

1. Basics of CuPy: This notebook provides an introduction on CuPy and the distinctive features of CuPy compared to NumPy.
2. Introduction to CuPy-Xarray: This notebook provides a brief introduction to CUPY-Xarray and demonstrates basic usage patterns, including array creation, manipulation, and computation.
3. Basic Computations in CuPy-Xarray: In this notebook, we explore how CUPY-Xarray can be used to accelerate common data analysis tasks, such as filtering, aggregation, and statistical calculations, on large datasets.
4. High-level Xarray Functions: CuPy vs. NumPy : In this notebook, we are going to explore high-level Xarray functions such as groupby, rolling mean, and weighted mean, and compared their execution times with traditional NumPy-based implementations.
5. Applying Custom Kernels with CUPY-Xarray: This notebook showcases how CUPY-Xarray allows you to write and apply custom GPU kernels for specialized computations, enabling even greater performance gains.

CuPy User Guide
Xarray User Guide
Cupy-Xarray Github NCAR GPU Workshop

• NCAR CISL/CSG Team
• ESDS Initiative