How can I check if "matlabpool" is running when using Parallel Computing Toolbox?

I am writing MATLAB scripts using the Parallel Computing Toolbox Release R2015b. Many of my scripts contain a call to "matlabpool" to start a parallel computing pool. However, sometimes the pool is already initiated and I will get an error due to a duplicate call.

How can I check if a parallel pool is already running, so I can avoid trying to start it twice?

Answer 1

[MathWorks Support Team]

Please note that the function "matlabpool" was deprecated in release R2016a. This answer contains three different methods to check the existence of a parallel pool and get its size depending on the MATLAB release being used.

For R2016a and later releases

Start a parallel pool with the command "parpool". You can then get the current parallel pool with the command "gcp" and the argument 'nocreate' to avoid creating a pool of none exists.

p = gcp('nocreate')

You can then check if the object 'p' exists, and how many workers it has assigned:

if isempty(p)
    % There is no parallel pool
    poolsize = 0;
else
    % There is a parallel pool of <p.NumWorkers> workers
    poolsize = p.NumWorkers
end

Function "parpool" is documented here: https://www.mathworks.com/help/parallel-computing/parpool.html

Function "gcp" is documented here: https://www.mathworks.com/help/parallel-computing/gcp.html

For releases between R2008b (7.7) and R2015b

You can type the following to check the size of "matlabpool":

poolsize = matlabpool('size')

The above command will return '0' if "matlabpool" is closed.

For releases earlier than R2008b:

The option 'size' was introduced in R2008b. For earlier releases, use the following function as a workaround:

function x = pool_size 
%POOL_SIZE - Return size of current MATLABPOOL 

session = com.mathworks.toolbox.distcomp.pmode.SessionFactory.getCurrentSession; 
if ~isempty( session ) && session.isSessionRunning() && session.isPoolManagerSession() 
    client = distcomp.getInteractiveObject(); 
    if strcmp( client.CurrentInteractiveType, 'matlabpool' ) 
        x = session.getLabs().getNumLabs(); 
    else 
        x = 0; 
    end 
else 
    x = 0; 
end 
end

[Walter Roberson]

@Riccardo Micci

What output do you see when you try these in R2022b ?

p = gcp('nocreate')

p = 

  0×0 Pool array with no properties.

isempty(p)

ans = logical
   1

pool = parpool

Starting parallel pool (parpool) using the 'Processes' profile ...
Connected to the parallel pool (number of workers: 2).

pool = 

 ProcessPool with properties: 

            Connected: true
           NumWorkers: 2
                 Busy: false
              Cluster: Processes (Local Cluster)
        AttachedFiles: {}
    AutoAddClientPath: true
            FileStore: [1x1 parallel.FileStore]
           ValueStore: [1x1 parallel.ValueStore]
          IdleTimeout: 30 minutes (30 minutes remaining)
          SpmdEnabled: true

p = gcp('nocreate')

p = 

 ProcessPool with properties: 

            Connected: true
           NumWorkers: 2
                 Busy: false
              Cluster: Processes (Local Cluster)
        AttachedFiles: {}
    AutoAddClientPath: true
            FileStore: [1x1 parallel.FileStore]
           ValueStore: [1x1 parallel.ValueStore]
          IdleTimeout: 30 minutes (30 minutes remaining)
          SpmdEnabled: true

isempty(p)

ans = logical
   0

Answer 2

[Seongsu Jeong]

>> isempty(gcp('nocreate'))

ans =

logical

   1

>> parpool(2)

Starting parallel pool (parpool) using the 'local' profile ... connected to 2 workers.

ans =

 Pool with properties: 

            Connected: true
           NumWorkers: 2
              Cluster: local
        AttachedFiles: {}
          IdleTimeout: 30 minutes (30 minutes remaining)
          SpmdEnabled: true

>> isempty(gcp('nocreate'))

ans =

logical

   0

>> delete(gcp('nocreate'))

Parallel pool using the 'local' profile is shutting down.

>> isempty(gcp('nocreate'))

ans =

logical

   1

>>