Another way instead of for loop

I have this code

 for i = 1:34
    for j = 1:i  % <-- Note the 1:i instead of 1:n
     s(i,j) = abs(z(i,ind(j))-z(j,ind(i)))/(z(i,ind(j))+z(j,ind(i)));
     s(j,i) = s(i,j)
    end
 end

firstly, i make a cluster to specific matrix(840000 x 840000) then i want to calculate previous code (with another way instead of for loop) to each point in the matrix

where 'ind' is cluster no that a point belong to, z is the distance from point to a cluster

Answer 1

[Guillaume]

The following should work. Basically, use ndgrid (or meshgrid) and sub2ind to compute all your indices at once:

[direct, indirect] = ndgrid(1:34, ind(1:34));
index1 = sub2ind(size(z), direct, indirect));
index2 = sub2ind(size(z), indirect, direct));
s = abs(z(index1) - z(index2)) ./ (z(index1) + z(index2));

I'm calculating the whole matrix at once instead of just the lower triangle as that will be faster anyway than flipping and copying the lower triangle.