parfor usage in parallel computing tool

Hi,

I am trying to implement parallel computing in MATLAB. The following code works,

A = zeros(10,5);

for k = 2:10
    parfor i =1:5
        A(k,i) = 1;
    end
end

However, if I change it a little,

A = zeros(10,5);

for k = 2:10
    parfor i =1:5
        A(k-1,i) = 1;
    end
end

an error returns

Error using parallel_computing_test (line 5)

Error: The variable A in a parfor cannot be classified.

See Parallel for Loops in MATLAB, "Solve Variable Classification Issues in parfor-Loops".

I already read the document, but I still do not understand what the problem is. Can anybody please help me?

Thank you in advance.

Answer 1

[Edric Ellis]

This is a documented restriction for sliced variables within parfor. The relevant doc page is here, and the relevant section is:

Form of Indexing. Within the first-level of indexing for a sliced variable, exactly one indexing expression is of the form i, i+k, i-k, or k+i. The index i is the loop variable and k is a scalar integer constant or a simple (non-indexed) broadcast variable. Every other indexing expression is a positive integer constant, a simple (non-indexed) broadcast variable, a nested for-loop index variable, colon, or end.

It's the sentence in bold at the end that is relevant. In your non-working case, the indexing expression k-1 does not meet the criteria. k is a broadcast variable, but the expression k-1 is not itself a broadcast variable. You could fix this like so:

A = zeros(10,5);
for k = 2:10
    kLess1 = k - 1;
    parfor i =1:5
        A(kLess1,i) = 1;
    end
end

because kLess1 is a broadcast variable.

[Yize Wang]

Thank you so much for the clear explanation!