3
votes

I have the following large, very inefficient loop.

P is a [2000 x 200 x 5] matrix
D is a [2000 x 200 x 5] matrix
S is a [200 x 1005] matrix
PS is a [2000 x 1000 x 5] matrix

I want to compute the following loop:

for k=1:2000
   for n=1:200
      for t=1:5
          P(k,n,t) = sum(S(n,t+1:t+1000) .* PS(k,1:1000,t));
      end
   end
end

Obviously this is very inefficient. I tried parfor, but I would rather a vectorized solution. I tried couple of things with bsxfun, but also never managed to get it working.

Thank you.

1
Could you a small sample data and the expected result,,please? It would help to verify any answers. - kkuilla
The data comes from a long code. I think for the purposes at hand any random data should do the trick: P = rand( 2000,200,5); D = rand( 2000,200,5); S = rand( 200,1005); PS = rand( 2000,1000,5); Thank you! - phdstudent
@volcompt please add data (or random data) to your question (i.e. not in the comments) and perhapd change the scale of the problem for the example or else replace the 1000, 200 and 5 with variables we can alter to work on a smaller problem. Also, you should post your bsxfun attempts and tell us where they go wrong. Lastly, your current code completely ignores the first column of S, is that correct? - Dan

1 Answers

6
votes

Here's an almost (almost because we still have a loop, but with only 5 iterations) vectorized approach using powerful matrix-multiplication -

out = zeros(2000,200,5);
for t=1:size(P,3) %// size(P,3) = 5
    out(:,:,t) = PS(:,:,t)*S(:,t+1:t+1000).';
end

Runtime tests and verify output -

%// Inputs
D = rand(2000,200,5);
S = rand(200,1005);
PS = rand(2000,1000,5);

disp('--------------------- No Matrix-mult-fun')
tic
P = zeros(2000,200,5);
for k=1:2000
   for n=1:200
      for t=1:5
          P(k,n,t) = sum(S(n,t+1:t+1000) .* PS(k,1:1000,t));
      end
   end
end
toc

disp('--------------------- Fun fun Matrix-mult-fun')
tic
out = zeros(2000,200,5);
for t=1:size(P,3) %// size(P,3) = 5
    out(:,:,t) = PS(:,:,t)*S(:,t+1:t+1000).';
end
toc

error_val = max(abs(P(:)-out(:)))

Output -

--------------------- No Matrix-mult-fun
Elapsed time is 70.223008 seconds.
--------------------- Fun fun Matrix-mult-fun
Elapsed time is 0.624308 seconds.
error_val =
     1.08e-12