How to speed up my code and improve/avoid for loop

Dear all, that's my piece of code that works correctly but too slow.
T=rand(100000,1)-0.3;
eTG=zeros(size(T)); %set eTG(1)=0 and preallocate
for i=2:length(T)
eTG(i)=eTG(i-1)+T(i-1);
eTG(i(eTG(i)>0))=0;
end
Preallocation doesn't help and "run and time" tool shows that string
eTG(i(eTG(i)>0))=0;
eats 83% of the total time. Please give me your advices how to optimize this code. Are there any ways to vectorize this code and avoid loops at all?

5 Comments

Given i is a scalar, what is
eTG(i(eTG(i)>0))=0
supposed to do?
Check if eTG variable is positive on every step and if yes, sets it to zero
Replace that with
eTG(i) = max(0, eTG(i));
Could you confirm that you want negative values to be left untouched, but positive values to be set to 0?
Yes, exactly. Using max function gives wrong results, Should be "min" I suppose. Thank you!
Dang, I got it wrong this time too. Yes, min(0, eTG(i)) if you want negative untouched and positive set to 0.

Sign in to comment.

Answers (1)

This code is 10 times faster (thanks to Walter Roberson)
T=rand(100000,1)-0.3;
eTG=zeros(size(T)); %set eTG(1)=0 and preallocate
for i=2:length(T)
eTG(i)=eTG(i-1)+T(i-1);
eTG(i) = min(0, eTG(i));
end

Categories

Find more on Parallel Computing Toolbox in Help Center and File Exchange

Products

Tags

Asked:

on 12 Oct 2017

Edited:

on 12 Oct 2017

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!