Introduction

The single program multiple data (spmd) language construct allows seamless interleaving of serial and parallel programming. The spmd statement lets you define a block of code to run simultaneously on multiple workers. Variables assigned inside the spmd statement on the workers allow direct access to their values from the client by reference via Composite objects.

This chapter explains some of the characteristics of spmd statements and Composite objects.

When to Use spmd

The “single program” aspect of spmd means that the identical code runs on multiple workers. You run one program in the MATLAB^® client, and those parts of it labeled as spmd blocks run on the workers. When the spmd block is complete, your program continues running in the client.

The “multiple data” aspect means that even though the spmd statement runs identical code on all workers, each worker can have different, unique data for that code. So multiple data sets can be accommodated by multiple workers.

Typical applications appropriate for spmd are those that require running simultaneous execution of a program on multiple data sets, when communication or synchronization is required between the workers. Some common cases are:

For more information, see Choose Between spmd, parfor, and parfeval.

Define an spmd Statement

The general form of an spmd statement is:

spmd
    <statements>
end

The block of code represented by <statements> executes in parallel simultaneously on all workers in the parallel pool. If you want to limit the execution to only a portion of these workers, specify exactly how many workers to run on:

spmd (n)
    <statements>
end

This statement requires that n workers run the spmd code. n must be less than or equal to the number of workers in the open parallel pool. If the pool is large enough, but n workers are not available, the statement waits until enough workers are available. If n is 0, the spmd statement uses no workers, and runs locally on the client, the same as if there were not a pool currently running.

You can specify a range for the number of workers:

spmd (m,n)
    <statements>
end

In this case, the spmd statement requires a minimum of m workers, and it uses a maximum of n workers.

If it is important to control the number of workers that execute your spmd statement, set the exact number in the cluster profile or with the spmd statement, rather than using a range.

For example, create a random matrix on three workers:

spmd (3)
    R = rand(4,4);
end

Unlike a parfor-loop, the workers used for an spmd statement each have a unique value for spmdIndex. This lets you specify code to be run on only certain workers, or to customize execution, usually for the purpose of accessing unique data.

For example, create different sized arrays depending on spmdIndex:

spmd (3)
    if spmdIndex==1 
        R = rand(9,9);
      else
        R = rand(4,4);
    end
end

Load unique data on each worker according to spmdIndex, and use the same function on each worker to compute a result from the data:

spmd (3)
    labdata = load(['datafile_' num2str(spmdIndex) '.ascii'])
    result = MyFunction(labdata)
end

The workers executing an spmd statement operate simultaneously and are aware of each other. As with a communicating job, you are allowed to directly control communications between the workers, transfer data between them, and use codistributed arrays among them.

For example, use a codistributed array in an spmd statement:

spmd (3)
    RR = rand(30, codistributor());
end

Each worker has a 30-by-10 segment of the codistributed array RR. For more information about codistributed arrays, see Working with Codistributed Arrays.

Display Output

When running an spmd statement on a parallel pool, all command-line output from the workers displays in the client Command Window. Because the workers are MATLAB sessions without displays, any graphical output (for example, figure windows) from the pool does not display at all.

MATLAB Path

All workers executing an spmd statement must have the same MATLAB search path as the client, so that they can execute any functions called in their common block of code. Therefore, whenever you use cd, addpath, or rmpath on the client, it also executes on all the workers, if possible. For more information, see the parpool reference page. When the workers are running on a different platform than the client, use the function pctRunOnAll to properly set the MATLAB path on all workers.

Error Handling

When an error occurs on a worker during the execution of an spmd statement, the error is reported to the client. The client tries to interrupt execution on all workers, and throws an error to the user.

Errors and warnings produced on workers are annotated with the worker ID (spmdIndex) and displayed in the client's Command Window in the order in which they are received by the MATLAB client.

The behavior of lastwarn is unspecified at the end of an spmd if used within its body.

spmd Limitations

a = 'a';
myFunction(a)

function X = myFunction(a)
name = inputname(1);
spmd
    X.(name) = spmdIndex;
end
X = [X{:}];
end
    

myFunction('a','b')

function myFunction(a,b)
nin = nargin;
spmd
    X = spmdIndex*nin;
end
end