parallel computing sig problem

Hi:

I have a for-loop, which is very long (over 500 line for example). now I'm able to optimize it into parallel computing.

the problem I'm having is that, I want to add a input as parallel sig: par_sig, to control if the code will do parallel computing. if par_sig=1, do parallel computing, if par_sig=0, do normal loop compution.

when this go to the code, it becomes:

    if par_sig==0
        for ...
            500 lines operation...
        end
    elseif par_sig==1
        parfor
            500 lines operation...
        end
    end

in this way the same 500 lines code are written twice in the code. if I revise somewhere I'll have to do twice also, which makes the operation very boring. is there anyway, or any trick, to write it only once?

Thanks!

Yu

Answer 1

[John D'Errico]

So, you simply cannot put those 500 lines of code inside a function? Or if not as a function, as a script m-file?

There is absolutely no need to copy the code and thus replicate it inline.

Essentially, you can just put it into a separate m-file. Call it my_code.m. Save that as a file. You would then call it essentially like this:

if par_sig==0
  for ...
    my_code
  end
elseif par_sig==1
  parfor
    my_code
  end
end

Now you can make any changes just once, editing the my_code function/script.

[Yu Li]

I have thought about this way, however, my code is looks like this:

    1000 lines code.
    for ...
        my_code      
    end

there is a lot of shared variables between 'my_code' and '1000 lines' code, it is very complex to put the 500 lines into 'my_code' directly.

however, thanks for your suggestion.

Yu