C# Job System Cookbook
This is a repo of examples I've written to learn how to use the C# job system to write systems at scale, here for reference and sharing.
The goal of this repo is making it clearer how you can structure your data, schedule your jobs, and use the results. So, the examples use easy to understand problems & algorithms.
This repo does not cover using C# jobs with the Entity Component System. Please see the official example repo for more on that.
Each example script has a corresponding scene where it's set up.
Job System Details
You can use the job system in Unity 2018.x right now. I recommend 2018.2 for these examples now.
Note: examples in this repo use
LateUpdate() as an easy way to handle completing jobs later than we schedule them, but in real code you might want to schedule the jobs early in
Update (using Script Execution Order maybe) so you can use the result later in the same frame.
All examples demonstrate the use of persistently-allocated job memory.
Process input from a webcam in real time using Burst-compiled jobs.
the job details are all in this file, and the above file is the main script.
This demo implements 5 different effects , all based around operating on a pixel only if it's color channel value is over some threshold
To change the color thresholds, select the
WebcamDisplay in the heirarchy of the example scene & check out the
Webcam Processing component. You can also change the scanline effect as well as select a webcam resolution that works for you there.
Modify all vertices & normals of a mesh in parallel every frame.
This is the most visually interesting example. Uses a more complex single job.
Modify all 20678 vertices of a mesh in parallel every frame, using Perlin noise & sin(time).
Uses a single job.
First determine velocities, then change positions based on those velocities.
Demonstrates using the TransformAccessArray, necessary for doing transform operations in jobs.
Vector3 and a
Bounds for intersection against a list of 10000
Demonstrates running 2 independent jobs.
Ray for intersection with a large
Bounds array in two steps.
Demonstrates reducing an array of checks to a smaller list, and using temporarily-allocated job memory.
Generates a cloud of 10000 points, then calculates magnitudes & normalizes the points.
Keijiro Takahashi has a great example of using the job system with ECS