malloc question

[Thoerin]

Could anyone explain why this program never seems to use more than ~90kB of memory (When checking with gnome-system-monitor, top, etc)? The pointer returned is not NULL when I pass a number less than the amount of free memory.

code:

#include <stdio.h> #include <stdlib.h> int main(int argc, char *argv[]){                         if(argc != 2)         {                          printf("Usage: leak [Number of MB to allocate]\n");                   return 1;         }                 unsigned int megabytes;                 sscanf(argv[1], "%u", &megabytes);         unsigned char *pointer = 0;                         pointer = malloc(megabytes * 1000000 * sizeof(unsigned char));         printf("Allocation complete.\n");         printf("%X", pointer);         int tmp = getchar();                    return 0;                }

[wtd]

Compiler optimization? Your compiler is realizing that you never actually use the memory allocated, so it doesn't bother to allocate any?

[Thoerin]

Hmmmm I tried compiling with -O0 before posting and it didnt help.

Just fixed it by setting to all zeros using memset, so I guess it was the optimizations after all.

Thanks for helping me out .

[btiffin]

Kinda. malloc usually reserves pages. They won't physically load until needed. Complex stuff, but read http://en.wikipedia.org/wiki/Paging for a little background.

The vmstat utility may show you a better picture of what is going on.

When you really care to see what a compiler may be doing in terms of optimization, use a -S switch (for gcc that is) and take a boo at the assembler output. The .s files rarely lie or hide anything from you; plus it can be a real eye opener exercise. I'm occasionally amazed at how little assembler some complex operations can actually take. Or vice versa, how some C source becomes a nearly unmanageable morass of assembler code to trace through by hand.

Cheers

[wtd]

Well yeah, when you're coding to god-awful complicated CPU architectures...