Version: 2018.4
Scheduling jobs
ParallelFor jobs

JobHandle and dependencies

When you call the Schedule method of a job it returns a JobHandle. You can use a JobHandle in your code as a dependency for other jobs. If a job depends on the results of another job, you can pass the first job’s JobHandle as a parameter to the second job’s Schedule method, like so:

JobHandle firstJobHandle = firstJob.Schedule();
secondJob.Schedule(firstJobHandle);

Combining dependencies

If a job has many dependencies, you can use the method JobHandle.CombineDependencies to merge them. CombineDependencies allows you to pass them onto the Schedule method.

NativeArray<JobHandle> handles = new NativeArray<JobHandle>(numJobs, Allocator.TempJob);

// Populate `handles` with `JobHandles` from multiple scheduled jobs...

JobHandle jh = JobHandle.CombineDependencies(handles);

Waiting for jobs in the main thread

Use JobHandle to force your code to wait in the main thread for your job to finish executing. To do this, call the method Complete on the JobHandle. At this point, you know the main thread can safely access the NativeContainer that the job was using.

Note: Jobs do not start executing when you schedule them. If you are waiting for the job in the main thread, and you need access to the NativeContainer data that the job is using, you can call the method JobHandle.Complete. This method flushes the jobs from the memory cache and starts the process of execution. Calling Complete on a JobHandle returns ownership of that job’s NativeContainer types to the main thread. You need to call Complete on a JobHandle to safely access those NativeContainer types from the main thread again. It is also possible to return ownership to the main thread by calling Complete on a JobHandle that is from a job dependency. For example, you can call Complete on jobA, or you can call Complete on jobB which depends on jobA. Both result in the NativeContainer types that are used by jobA being safe to access on the main thread after the call to Complete.

Otherwise, if you don’t need access to the data, you need to explicity flush the batch. To do this, call the static method JobHandle.ScheduleBatchedJobs. Note that calling this method can negatively impact performance.

An example of multiple jobs and dependencies

Job code:

// Job adding two floating point values together
public struct MyJob : IJob
{
    public float a;
    public float b;
    public NativeArray<float> result;

    public void Execute()
    {
        result[0] = a + b;
    }
}

// Job adding one to a value
public struct AddOneJob : IJob
{
    public NativeArray<float> result;
    
    public void Execute()
    {
        result[0] = result[0] + 1;
    }
}

Main thread code:

// Create a native array of a single float to store the result in. This example waits for the job to complete
NativeArray<float> result = new NativeArray<float>(1, Allocator.TempJob);

// Setup the data for job #1
MyJob jobData = new MyJob();
jobData.a = 10;
jobData.b = 10;
jobData.result = result;

// Schedule job #1
JobHandle firstHandle = jobData.Schedule();

// Setup the data for job #2
AddOneJob incJobData = new AddOneJob();
incJobData.result = result;

// Schedule job #2
JobHandle secondHandle = incJobData.Schedule(firstHandle);

// Wait for job #2 to complete
secondHandle.Complete();

// All copies of the NativeArray point to the same memory, you can access the result in "your" copy of the NativeArray
float aPlusB = result[0];

// Free the memory allocated by the result array
result.Dispose();

Scheduling jobs
ParallelFor jobs
Copyright © 2023 Unity Technologies
优美缔软件(上海)有限公司 版权所有
"Unity"、Unity 徽标及其他 Unity 商标是 Unity Technologies 或其附属机构在美国及其他地区的商标或注册商标。其他名称或品牌是其各自所有者的商标。
公安部备案号:
31010902002961