Parallel Computing Question for For loop switch to parfor

I am trying to find the nearest phase value to time intervals. This is the loop I have used. The matrix is large and I want to use parallel computing function 'parfor', but cannot due to having embedded functions within a function. I somewhat understand that they need things to be as simplistic as possible, but is there a way to fix this so that I can use 'parfor' function? I do not fully understand why the parallel processing has the limitations, so I am not exactly sure how to fix it.

I attached matrices for 'timeintervals' and 'phase values' if anyone needs them.

Any help would be greatly appreciated!

for i=1:length(timeintervals(:,1))
  for ii=1:length(timeintervals)
  [~,ix]=min(abs(timeintervals(i,ii)-phasevalues(:,2)));  % find nearest location
  out(i,ii)=phasevalues(ix);   
  end   
end

[Matt J]

but cannot due to having embedded functions within a function.

I don't see any function calls in your code.

Answer 1

[Matt J]

Why not as follows?

 col1=phasevalues(:,1);
 col2=phasevalues(:,2); %as separate variable, repeated memory allocation is avoided

 parfor j=1:size(timeintervals,2)
   [~,ix]=min( abs(timeintervals(:,j).' - col2 ) );  % find nearest location
   out(:,j)=col1(ix);   
 end

[Systematically Neural]

Works great, thank you!