MATLAB Answers

command in Matlab equivalent to "forall " in fortran

17 views (last 30 days)
smabtahi
smabtahi on 24 May 2012
Hi Is there any command in matlab that can simpilfy such loop:
for i=1:10 A(i)=i^2 end
as I know in fortran it can be written anr run very fast as FORALL(i=1,10) A(i)=i**2
I hope similar thing can be found in Matlab too.

  0 Comments

Sign in to comment.

Accepted Answer

Daniel Shub
Daniel Shub on 24 May 2012
The JIT accelerator that was introduced 10 years ago has sped up loops substantially, it is an unfortunate myth that MATLAB loops are slow. Vectorization can speed up some loops, but it can also slow things down. The ability to use parfor loops can also speed up things (and allow scaling) that is not easy with vectorized code.
For you case of wanting a user-defined function to be vectorized, the easy answer is to go back to the function (jmat_TM) and modify it to handle vector inputs.

  2 Comments

smabtahi
smabtahi on 24 May 2012
Thanks Daniel and Jan
Then it seems that I have no other choise than vectorizing jmat_TM.
Daniel Shub
Daniel Shub on 24 May 2012
Or you could just stick with your loop. There is no guarantee that vectorizing will speed things up, although it will make your code for jmat_TM much more MATLAB like.

Sign in to comment.

More Answers (1)

Wayne King
Wayne King on 24 May 2012
A = 1:10;
A = A.^2;

  2 Comments

smabtahi
smabtahi on 24 May 2012
Thank you Wayne.
of course the power function in pre-defined in matlab for whole matrixes, but consider a user defined function, for instance this loops:
for id=1:10
for jmd=1:100;
[Jsen(id,jmd)]=jmat_TM(id,jmd);
end
end
and jmat_TM is a user-defined complicated function. I want to get rid of using those loops.
Jan
Jan on 24 May 2012
You can omit the square brackest around the output here.
You need either the loops or rewrite the user defined function, that it accepts matrices as inputs. Therefore Wayne's suggestion is fine.

Sign in to comment.

Tags

Community Treasure Hunt

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

Start Hunting!