Variable loop depth - Make into Matrix?

Hello,

I have a situation where I am trying to make a comb function for each of a series of variables. The code I have thus far is as follows

for row_delay = 1:num_row_dealays
   for hole delay = 1: num_hole_delays
     for row_num = 1:num_rows
       for hole_num = 1:num_holes
         if (design_array(row_num,hole_num)) % Array has either 1 or 0
            ftime = hole_delays(hole_delay) * hole_num + row_delays(row_delay) * (row_num-1); % hole_delays predefined
            comb_array(row_delay,hole_delay,ftime) = comb_array(row_delay,hole_delay) + 1;
         end
       end
    end
  end
end

This works, but now I need to have a variety of depths (number of different "row_delays" arrays).

Can this kind of loop be made into a matrix operation. Clearly, the "comb_array" will have to have different arguments, like "row_delay1", "row_delay2", and the number of them will vary, so simplifying this would be great!

Thanks! Any suggestions are welcome.

Doug Anderson

[Matt Sprague]

You can use logical indexing to replace for loops. It's hard to fully understand what is happening in the "comb_array" line since you are referring to the variable on the left side of the equation with 3 dimensions but on the right side only with 2 dimensions. However, the optimized code should look something along the lines of:

row_delay = 1:num_row_dealays;
hole_delay = 1: num_hole_delays;
row_num = 1:num_rows;
hole_num = 1:num_holes;
%
x = design_array(row_num,hole_num); % Array has either 1 or 0
ftime = hole_delays(hole_delay(x)) .* hole_num(x) + row_delays(row_delay(x)) .* (row_num(x)-1); % hole_delays predefined
%
comb_array(row_delay(x),hole_delay(x),ftime) = comb_array(row_delay(x),hole_delay(x)) + 1;