loop optimisation in c++

Question

0.00/5 (No votes)

See more:

How can i further optimise the for loop in the codes below?is there also a way to avoid declarations inside the loop ie int x0,x1, etc?

C#

void tom::add(void* btr)
{

      __declspec(align(16))short* b =(short*)btr;
     int j;

    
        
        for(j = 0; j < 16; j += 4)
        {
            /// 1st stage transform.
            int x0 = (int)(b[j]     + b[j+3]);
            int x3 = (int)(b[j]     - b[j+3]);
            int x1 = (int)(b[j+1] + b[j+2]);
            int x2 = (int)(b[j+1] - b[j+2]);

            /// 2nd stage transform.
            b[j]        = (short)(x0 + x1);
            b[j+2]  = (short)(x0 - x1);
            b[j+1]  = (short)(x2 + (x3 << 1));
            b[j+3]  = (short)(x3 - (x2 << 1));
        }

Posted 29-Aug-10 1:28am

SMART LUBOBYA

Add a Solution

2 solutions

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

#realJSOP · Answer 1 · 2010-08-29T01:57:00

Solution 1

Yeah, put the declarations outside the loop.

You'd also save a lot of time by not casting in the loop. If you need an int for the calculations, cast the incoming parameter to an int before you start the loop, do the loop, and then cast the result back to a short after the loop.

Posted 29-Aug-10 1:57am

#realJSOP

Niklas L · Answer 2 · 2010-08-29T08:50:00

Declaring the int's inside the loop does not affect performance. The allocation for that object is done on the stack when you enter the function. Also an int has no constructor, which is the crucial part. Declaring an int does not generate any code.

Objects having non-trivial constructors is another issue, and should be given special concern when dealing with loops that requires performance.

Also, watch out for overflow in statements like this

(int)(b[j] + b[j+3]);

If you have nothing better to do, you could also unroll the loop.

Edit:
There is something else bothering me with your code. Why do you use a void* as argument to your function if you only accept a short* implementing it? Would it not be better to have a short* as argument and let the caller, who probably knows your void pointers content a lot better than the add() function, do the conversion?