In C,

Suppose you want to declare a variable, and say you need some initial numberic value in it, and then you are going to use that variable later, like:

int var1;
var1=0;
var1=var1+7;

int var1=0;
var1=var1+7;

8/30/2005 6:03:00 PM

for a good compiler - same

8/30/2005 6:09:23 PM

I don't know for sure, but I would guess they compile to the same thing.

8/30/2005 6:09:36 PM

Write your code so it is as readable as possible, document well, and get a good compiler.

8/30/2005 6:11:00 PM

trying to make optimizations like these are idiotic unless your time is not very valuable. Most of your time should be focused on design and algorithmic optimizations.

8/30/2005 8:01:42 PM

yea, this question is only relevant if you're writing assembly - and even then, i'd hope the assembler would assemble it the same either way.

8/30/2005 8:15:05 PM

Both methods will reduce to something like this (probably exactly like this, at least for msvc++):

mov dword ptr [var1],0
mov eax,dword ptr [var1]
add eax,7
mov dword ptr [var1],eax

^^ While that is probably true for plain old C, in C++ you have to be very careful with default versus copy constructors. For example, if you have a class that is expensive to construct but lightweight in terms of copy construction and assignment, then tracking down the places where you do:

someClass obj;
obj = obj2;

and changing it to:

someClass obj = obj2;

could potentially give you huge performance wins.

8/30/2005 8:21:13 PM

FORTRAN bitches

8/30/2005 11:06:08 PM

int var1=0;
var1+=7;

[Edited on August 31, 2005 at 12:00 AM. Reason : the important stuff has been said, that's just how I'd write it]

8/30/2005 11:59:19 PM

int var1 = 7;

damn

8/31/2005 10:49:23 AM

#define var1 7

I win.

8/31/2005 12:56:15 PM

^ he wins

8/31/2005 2:42:52 PM

7;
i win

8/31/2005 10:52:18 PM