Ray "Crash Course" Overview¶
© 2019-2021, Anyscale. All Rights Reserved
Join Us at Ray Summit 2020!¶
Join us for the free Ray Summit 2021 virtual conference, June 22-24, 2021. We have an amazing lineup of luminar keynote speakers and breakout sessions on the Ray ecosystem, third-party Ray libraries, and applications of Ray in the real world.
About This Tutorial¶
This tutorial, part of Anyscale Academy, provides a "just the facts" introduction to Ray, the system for scaling your applications from a laptop to a cluster. Here, we start with a hands on exploration of the core Ray API for distributed work loads:
- Ray tasks: Distributed execution of stateless Python functions.
- Ray actors: Distributed execution of stateful Python classes.
Subsequent lessons explore the origin and goals for Ray, Ray replacement versions of several popular multiprocessing APIs, a Ray parallel iterator API, the rest of the core Ray API (documentation), and working with Ray clusters using the ray command-line interface (CLI). Pick among those lessons according to your interests.
Tips:
- See the instructions in the README for setting up your environment to use this tutorial
- Having problems? See Troubleshooting, Tips, and Tricks
- Go here for an overview of all tutorials.
Getting Help¶
- The #tutorial channel on the Ray Slack. Click here to join.
Find an issue? Please report it!
"Crash Course" Tutorial Guide¶
| # | Lesson (Notebook) | Description |
|---|---|---|
| 00 | Overview | A table of contents for this tutorial. |
| 01 | Ray Tasks | Understanding how Ray converts normal Python functions into distributed stateless tasks. |
| 02 | Ray Actors | Understanding how Ray converts normal Python classes into distributed, stateful actors. |
| 03 | Why Ray? | Now that you've seen Ray in action, take a step back and explore the origin and motivations for Ray, and Ray's growing ecosystem of libraries and tools. |
| 04 | Ray Multiprocessing | Ray drop-in replacements for common parallelism APIs and integration with asyncio. |
| 05 | Ray Parallel Iterators | As of Ray 1.7, ray.util.iter has been deprecated. Don't run this notebook. A parallel iterator API for simple data ingestion and processing. It can be thought of as syntactic sugar around Ray actors and ray.wait loops. |
| 06 | Exploring Ray API Calls | The Ray API has other API calls for more advanced scenarios, which are surveyed in this optional lesson. Keyword arguments you can pass to the API calls already learned are explored. |
| 07 | Running Ray Clusters | A brief look at the Ray CLI commands for running Ray clusters. |
In addition, exercise solutions for this tutorial can be found here.