When you schedule a job there can only be one job doing one task. However, there will be times where you need to perform the same operation on a lot of objects. To do this, use a ParallelFor job type, which inherits from IJobParallelFor
.
A ParallelFor job uses a NativeArray
of data to act on as its data source. ParallelFor jobs run across multiple CPU cores. There’s one job per core, with each handling a subset of the workload.
IJobParallelFor
behaves like IJob
, but instead of a single Execute
method, it invokes the Execute
method once per item in the data source. There’s also an integer parameter index in the Execute
method, which you can use to access and operate on a single element of the data source within the job implementation.
The following is an example of a ParallelFor job definition:
struct IncrementByDeltaTimeJob: IJobParallelFor
{
public NativeArray<float> values;
public float deltaTime;
public void Execute (int index)
{
float temp = values[index];
temp += deltaTime;
values[index] = temp;
}
}
To schedule a ParallelFor
job, you must specify the length of the NativeArray
data source that you want to split. The job system doesn’t know which NativeArray
you want to use as the data source if there are several in the struct. The length also tells the job system how many Execute
methods to expect.
In Unity’s native code, the scheduling of ParallelFor
jobs is more complicated. When Unity schedules a ParallelFor
job, the job system divides the work into batches to distribute between cores. Each batch contains a subset of Execute
methods. The job system then schedules one job in Unity’s native job system per CPU core and passes that native job to the batches to complete.
When a native job completes its batches before others, it steals remaining batches from the other native jobs. It only steals half of a native job’s remaining batches at a time, to ensure cache locality.
To optimize the process, you need to specify a batch count. The batch count controls how many jobs you get, and how fine-grained the redistribution of work between threads is. Having a low batch count, such as 1, gives you an even distribution of work between threads. However, it comes with some overhead, so sometimes it’s better to increase the batch count. Starting at 1 and increasing the batch count until there are negligible performance gains is a good strategy.
The following is an example of scheduling a ParallelFor job
Job code:
// 将两个浮点值相加的作业
public struct MyParallelJob : IJobParallelFor
{
[ReadOnly]
public NativeArray<float> a;
[ReadOnly]
public NativeArray<float> b;
public NativeArray<float> result;
public void Execute(int i)
{
result[i] = a[i] + b[i];
}
}
Main thread code:
NativeArray<float> a = new NativeArray<float>(2, Allocator.TempJob);
NativeArray<float> b = new NativeArray<float>(2, Allocator.TempJob);
NativeArray<float> result = new NativeArray<float>(2, Allocator.TempJob);
a[0] = 1.1;
b[0] = 2.2;
a[1] = 3.3;
b[1] = 4.4;
MyParallelJob jobData = new MyParallelJob();
jobData.a = a;
jobData.b = b;
jobData.result = result;
// 调度作业,为结果数组中的每个索引执行一个 Execute 方法,且每个处理批次只处理一项
JobHandle handle = jobData.Schedule(result.Length, 1);
// 等待作业完成
handle.Complete();
// 释放数组分配的内存
a.Dispose();
b.Dispose();
result.Dispose();
A ParallelForTransform
job is another type of ParallelFor
job designed specifically for operating on Transforms. It’s useful for working on Transform operations from jobs efficiently. For more information, see the ParallelForTransform API documentation.
Did you find this page useful? Please give it a rating:
Thanks for rating this page!
What kind of problem would you like to report?
Thanks for letting us know! This page has been marked for review based on your feedback.
If you have time, you can provide more information to help us fix the problem faster.
Provide more information
You've told us this page needs code samples. If you'd like to help us further, you could provide a code sample, or tell us about what kind of code sample you'd like to see:
You've told us there are code samples on this page which don't work. If you know how to fix it, or have something better we could use instead, please let us know:
You've told us there is information missing from this page. Please tell us more about what's missing:
You've told us there is incorrect information on this page. If you know what we should change to make it correct, please tell us:
You've told us this page has unclear or confusing information. Please tell us more about what you found unclear or confusing, or let us know how we could make it clearer:
You've told us there is a spelling or grammar error on this page. Please tell us what's wrong:
You've told us this page has a problem. Please tell us more about what's wrong:
Thank you for helping to make the Unity documentation better!
Your feedback has been submitted as a ticket for our documentation team to review.
We are not able to reply to every ticket submitted.