Oct 18 2022

GPU-Accelerated Computing with Python

October 18, 2022

1:00 PM - 4:00 PM

Description

In this workshop, you'll get hands-on experience accelerating Python codes with NVIDIA GPUs. We will utilize code samples in three main categories to introduce you to Python GPU accelerated computing. First, we will explore drop-in replacements for SciPy and NumPy code through the CuPy library. Next we’ll cover NVIDIA RAPIDS, which provides GPU acceleration for end-to-end data science workloads. Finally we'll cover Numba, which gives you the flexibility to write custom accelerated code without leaving the Python language. We'll finish with an end-to-end example that incorporates all the tools introduced to tackle a geospatial problem. By the end of the workshop, you'll have the skills to start accelerating your own Python codes with NVIDIA GPUs!

Learning Objectives

By participating in this workshop, you’ll learn how to:

  • Learn to accelerate existing Python applications on GPU
  • Introduce CuPy and RAPIDS for drop-in GPU-accelerated replacements for Numpy and Scikit learn
  • Introduce Numba for writing custom GPU accelerated functions

Prerequisites Basic familiarity with Python

  • Familiarity with Numpy, and/or Scikit-learn is beneficial

Tools, Libraries, and Frameworks Used

  • Numba
  • CuPy
  • RAPIDS

Course Length: 3 hours

This workshop speaker will be Kristopher Keipert, senior solutions architect at NVIDIA.

Author(s):  Kristopher Keipert and Zoe Ryan

Registration

Registrations for this training is closed.

Contact

Ana Marija Sokovic

Date posted

Sep 2, 2022

Date updated

Oct 14, 2022