parfor supply error (no error with for)

Problem:

When I run with for, everything works, when I use parfor at either the sub or the i loop level, I get a "parfor supply" error.

Description:

Input: I have a 4-dimensional dataset (15x[fixed size]x[variable size]x[variable size]).

Process: I want to perform the same functions on all of the 15 3-d elements, with four outputs each.

Output: I have tried to implement this by creating a 15x1 output cell, wherein each cell has a structure with four fields (a [2d matrix], b [3d matrix], c [3d matrix], [4d matrix]) or in the following way (4 separate matrices), without success.

load('data.mat'); %data is a 15 x 1535 x ... x ... matrix
permi=100; %number of permutations
subseti=1:5:size(data,2); %select every 5th element for extra processing

parfor sub=1:nsub 

      subdata=squeeze(data(sub,:,:,:)); %extract relevant 3-d data

      for i=1:size(data,2) 

          microdata=squeeze(subdata(i,:,:)); %data for the (i)th element

          a(sub,i)=foo(microdata); %process 1 on current (i)th element               
          b(sub,i,:)=foo2(microdata); %process 2 on current (i)th element       
          c(sub,i,:)=foo3(microdata); %process 2 on current (i)th element

          if find(i==subseti) 

              for permi=1:permutei             

                  d(sub,permi,i)=foo4(microdata); 

              end;
          end;           
      end;       
  end;

[Steven Lord]

What is the FULL text of the error you receive when you run that code? Post ALL of the red text; don't omit or paraphrase anything.

[Alexandra List]

script.m>(parfor supply)

Answer 1

[Sean de Wolski]

My guess is that parfor is unable to determine what variables it needs due to how you called load. Instead call load with an output and grab each of the variables from it; e.g:

S = load('data.mat');
x = S.x;
y = S.y;