how can I improve this code???!!!
2 vues (au cours des 30 derniers jours)
Afficher commentaires plus anciens
hi , I write a script and in this script below code take time than other codes in my script.
I need your knowledge to know how can I improve this script. (I know it takes only 0.11 sec to run but in large scale inputs it takes too long and also I want to learn that how can I write a script efficient.)
margin=zeros;
l=1;
C_in= 3405;
for i=1:1824
for j=1:10
temp=COPT(i,3)*COPT_load(j,2);
if temp>d
%this line takes time %from here
margin(l,1)=COPT(i,2)-COPT_load(j,1);
%to here
margin(l,2)=temp;
l=l+1;
end
end
end
So how can I improve my code ???
2 commentaires
the cyclist
le 13 Nov 2019
Can you upload a *.mat file with the data, so that we can test on your actual data?
Réponse acceptée
Daniel M
le 13 Nov 2019
Modifié(e) : Daniel M
le 13 Nov 2019
You should try preallocating margin before the loop. You have only set the value to one element, of zero. Since it's not clear how big margin will be at the end of your loops, you should preallocate it to be the maximum size (1824*10 = 18240). Then at the end, only keep the first 1:l (L) rows using
% after your loops
margin = margin(1:l,:);
Let us know if it's still slow after that, and include a mat file so we can test.
4 commentaires
Steven Lord
le 14 Nov 2019
You don't need to know the exact size of the array you want to preallocate when you write the code, if you can determine or calculate the size of the array from information available before you enter the loop(s). I'm going to guess the limits on i and j are related to the size of COPT?
assert(ismatrix(COPT)) % rather than a 3-D or N-D (N > 2) array
sz = size(COPT);
% Preallocate A to be the same size as COPT
A = zeros(sz);
for whichrow = 1:sz(1)
for whichcol = 1:sz(2)
% Do stuff with A(whichrow, whichcol) and COPT(whichrow, whichcol)
end
end
I have no idea how large COPT will be when you run that code, but I do know A will start off the same size as COPT.
Daniel M
le 14 Nov 2019
Modifié(e) : Daniel M
le 14 Nov 2019
Incremental growth of arrays and dynamic reallocation are tricky things in MATLAB (this is when you cannot know the size that an array will be before the loop occurs). You may be interested in the utility growdata on the File Exchange, where John D'Errico talks about and implements solutions to this problem:
Plus de réponses (0)
Voir également
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!