GitHub

Basic Scheduling

Learn the fundamentals of job scheduling in Gears. This tutorial covers the core concepts you need to understand how jobs are scheduled and executed.

The Scheduling Model

Gears schedules jobs based on the current time as provided by the active clock and scheduler. Jobs are scheduled to run at specific times, whenever an event is triggered, or just as soon as possible. The scheduler determines when each job should execute according to its configuration.

Note

Before we dive into the basics of scheduling, let's cover the core concepts of clocks and schedulers. These concepts will be covered in more detail in the Clocks and Time and Schedulers tutorials. For now, it is enough to know that:

  • A Scheduler is responsible for keeping a registry of jobs, and calling jobs when they are due.
  • A Clock is responsible for providing the current time. A Scheduler uses a Clock to determine when jobs should be executed.

Core Functions

every() - Schedule Jobs

The every() function is your primary tool for scheduling jobs:

# Schedule a job every 100 milliseconds
every(100ms) do dt
    println("Executed after $dt seconds")
end

The dt parameter represents the time elapsed since the last execution. Here, we have configured a job to run on a time interval, and the scheduler will execute the job every 100 milliseconds. A job that runs on a time interval is called a timed job and always takes a dt parameter, so the body can depend on the time elapsed since the last execution.

for_next() - Run the Scheduler

Use for_next() to run the scheduler for a specific duration:

# Run the scheduler for 500 milliseconds
for_next(500ms) do
    update!()
end

update!() - Process Jobs

update!() advances time and executes all jobs that are due:

update!()  # Process all scheduled jobs

A Complete Example

Let's build a simple simulation that demonstrates basic scheduling:

using Gears

# Schedule a sensor reading every 50ms
every(50ms) do dt
    reading = rand() * 100
    println("Sensor reading: $(round(reading, digits=2))")
end

# Schedule a status report every 180ms
every(180ms) do dt
    println("Status report at $(now(global_clock()))")
end

# Run the simulation for 1 second
for_next(1.0s) do
    update!()
end

Understanding the Output

Notice that:

  • The sensor runs every 50ms (4 times per 200ms)
  • The status report runs every 180ms
  • Both jobs run independently

Key Concepts

  1. Independent Jobs: Each every() call creates a separate job that runs independently
  2. Synchronous Execution: All jobs due at a given time are executed before time advances

Next Steps

Now that you understand basic scheduling, you're ready to learn about: