spmdBarrier

Synchronize workers in spmd block

Since R2022b

Syntax

spmdBarrier

Description

spmdBarrier stops all workers in the current spmd block or communicating job from executing code until every worker calls spmdBarrier.

When you offload computations using parfor and parfeval, only one worker at a time runs each computation. These workers are independent and do not communicate with each other. If you apply spmdBarrier to these workers, the function has no effect.

Use spmdBarrier to synchronize workers, for example, when workers use shared resources such as a file handle.

If only one worker is running on the current spmd block, execution continues immediately. To determine the number of workers running the current spmd block, use the spmdSize function. The spmdSize function returns a value of 1 outside of an spmd block or communicating job.

Examples

Synchronize Workers in spmd Block

This example shows how to use spmdBarrier to synchronize workers in an spmd block.

Create a parallel pool with four workers.

parpool(4);

When you execute an spmd block after creating a parallel pool, by default all available workers in the pool run the code inside the spmd block.

Create an spmd block. Pause each worker for a random amount of time to simulate some computationally expensive work. Use tic and toc to time the execution on each worker.

spmd
    tic
    pause(5*rand);
    toc
end

Worker 2: 
  Elapsed time is 0.702969 seconds.
Worker 3: 
  Elapsed time is 1.807292 seconds.
Worker 1: 
  Elapsed time is 4.651690 seconds.
Worker 4: 
  Elapsed time is 4.694443 seconds.

To synchronize the workers after each worker runs pause, use spmdBarrier. All the workers wait for the slowest worker to finish its computation. The elapsed time on each worker is now the same, except for small numerical noise.

spmd
    tic
    pause(5*rand);
    spmdBarrier;
    toc
end

Worker 1: 
  Elapsed time is 4.758529 seconds.
Worker 2: 
  Elapsed time is 4.758529 seconds.
Worker 3: 
  Elapsed time is 4.743785 seconds.
Worker 4: 
  Elapsed time is 4.743739 seconds.

Extended Capabilities

Thread-Based Environment

Run code in the background using MATLAB® backgroundPool or accelerate code with Parallel Computing Toolbox™ ThreadPool.

This function fully supports thread-based environments. For more information, see Run MATLAB Functions in Thread-Based Environment.

Version History

Introduced in R2022b