Parallel execution using parfeval does not modify input handle object as expected

I'm trying to implement parallelism - I want to run my function in parallel, and it's supposed to modify its input OOP object asynchronously.

Here is a simplified test implementation that runs, but the input object is NOT modified by the function calls.

if isempty(gcp('nocreate'))
    % Start a new parallel pool
    parpool();
end

futs= Futures();
nodeCount = NodeCount(1);


for i=1:5
    f=parfeval(@Testparffun,0,nodeCount);
    futs.append(f)
end

If the parfeval is replaced by the normal function call, it works as expected. With parfeval, the code completes without errors, but nodeCount is not modified .

Here is also the code for Testparffun and NodeCount:

function  Testparffun(counter)

    counter.increment;

    pause(5);

end

And NodeCount:

classdef NodeCount < handle
    %NODECOUNT Summary of this class goes here
    %   Detailed explanation goes here
    
    properties
        count
    end
    
    methods
        function obj =  NodeCount(inputArg1)
            %NODECOUNT Construct an instance of this class
            %   Detailed explanation goes here
            obj.count =inputArg1 ;
        end
        
        function   increment(obj)
            %METHOD1 Summary of this method goes here
            %   Detailed explanation goes here
            obj.count = obj.count + 1;
        end
    end
end

The goal of this test would be that each async call increments the value of nodeCount.count by 1, but it does nothing. I understand that I would need to implement some locks to avoid race conditions, etc. but even if the loop makes just one call, it still doesn't affect the object. Any ideas?

Answer 1

[Steven Lord]

See this documentation page for an explanation and a description of how I think you can achieve the behavior you're looking for.