While Loop vs For Loop, which is more memory efficient!

Question

Now that I’m used to programming in conditions of limited memory, one question that I don’t have an answer to is: what is more from the point of view of memory: - for (;;) or while ()? Or are they interchangeable equally? Also comment on the effectiveness, if any!

Answer 1

There is no difference in memory usage between the for and while programming language constructs. This compiler can optimize one or the other better, but you should not worry about it. With limited memory, your data structures and system design will be most important. Poor system design may require more memory.

Answer 2

The only way to make sure that you need to test it with your compiler for your platform. But I suppose that any sufficiently advanced optimizing C compiler in most cases optimizes them for the same thing.

Answer 3

 for(init;cond;inc){ //... } 

(almost) just like

 init; while(cond){ //... inc; } 

the differences are that continue; in for also does inc; but you have to do it manually during

Variables declared in the init; statement are limited by the for parameter (it introduces a new scope), however this is not the case when

this simplifies the use of loops in several cases (for example, iterating over an array), however, the compiled code is likely to be the same, so the only reason for choosing one over the other is readability

note I am talking mainly about java. I don't know the features in c well enough to comment on this.

Answer 4

Well, it depends on the compiler ... but think of it this way:

 while(x){ x++; } 

uses a variable declared outside the scope of the loop, i.e. you are using data already made for the loop.

the for loop initializes variables only for the loop inside the loop.

 for(int x = 0; x > y; x++{ } 

So, if you have a variable just sitting around that has a value that you can use, most likely the while loop will (depending on the compiler) be better for this situation. But if you needed to make variables only for a loop, I would use a for loop.

Hope this helps.

Answer 5

I did some testing in program c using the clock and the results were inconclusive .. so I wrote the code

 for(int i=0;i< 100000000; i++){ g+=2; } 

and code

 int i2 = 0; while(i2 < 100000000){ g+=2; i2++; } 

when compiling in dev C ++ using gcc, the assembly breaks into exactly the same bytecode ...

 CPU Disasm; Address Hex dump Command Comments 0040160C |. C745 E0 00000 MOV DWORD PTR SS:[LOCAL.8],0 00401613 |> 817D E0 FFE0F /CMP DWORD PTR SS:[LOCAL.8],5F5E0FF 0040161A |. 7F 0D |JG SHORT 00401629 0040161C |. 8D45 E4 |LEA EAX,[LOCAL.7] 0040161F |. 8300 02 |ADD DWORD PTR DS:[EAX],2 00401622 |. 8D45 E0 |LEA EAX,[LOCAL.8] 00401625 |. FF00 |INC DWORD PTR DS:[EAX] 00401627 |.^ EB EA \JMP SHORT 00401613 

both sets of code deal with this exact assembly node so you can see that they are equally good. this in combination with watch tests always returns as one faster once, and the other faster, at another time proves it ...

if you want you can find your code in a disassembler and optimize it manually;)

for example, these cycles can be represented as follows

 CPU Disasm Address Hex dump Command Comments 00401492 |. 60 PUSHAD 00401493 |. 31DB XOR EBX,EBX 00401495 |. 31C0 XOR EAX,EAX 00401497 |> 81FB FFE0F505 /CMP EBX,5F5E0FF 0040149D |. 7F 05 |JG SHORT 004014A4 0040149F |. 40 |INC EAX 004014A0 |. 40 |INC EAX 004014A1 |. 43 |INC EBX 004014A2 |.^ EB F3 \JMP SHORT 00401497 004014A4 |> 8945 E4 MOV DWORD PTR SS:[LOCAL.7],EAX 004014A7 |. 61 POPAD 

you can change two INC EAX to ADD EAX, 2, but this uses 1 more byte, which I did not want to lose :)

this optimization is about twice as fast :)