Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Here is a really dumb example:

  $ cat bigbuf.c 

  #include <stdlib.h>
  #include <stdio.h>
  #include <string.h>

  #define DEFINITELY_BIG_ENOUGH 2U*1024*1024*1024
  int main(int nargs, char ** args)
  {

    char * definitely_big_enough = malloc(DEFINITELY_BIG_ENOUGH);

    if (nargs > 1)
    {
      memset(definitely_big_enough,0, DEFINITELY_BIG_ENOUGH);
    }

    sprintf(definitely_big_enough,"%s",args[0]);

    return 0;
  }
  
  $ /usr/bin/time ./bigbuf 
  0.00user 0.00system 0:00.00elapsed 100%CPU (0avgtext+0avgdata 1356maxresident)k
  0inputs+0outputs (0major+80minor)pagefaults 0swaps

  $ /usr/bin/time ./bigbuf 1 
  0.05user 1.25system 0:01.31elapsed 99%CPU (0avgtext+0avgdata 2098468maxresident)k
  0inputs+0outputs (0major+524369minor)pagefaults 0swaps
YMMV on different operating systems. Of course this is a program only an idiot would write, but things like caches are often significantly bigger than the median case, especially on Linux where you know there is overcommit.


A non-idiot would use calloc(DEFINITELY_BIG_ENOUGH), and that will likely erase the difference because the impl will be able to rely on mmap(ANONYMOUS) creating zero pages for such a large allocation. A more realistic test would be to have a large number of small allocations that get calloc'd and then free'd repeatedly, because a) free'ing a small allocation will not free the underlying page and thus reallocation will not be able to rely on it already being zero, and b) zeroing small allocations doesn't amortize the cost of zeroing as well as zeroing large allocations does.


c) SIMD won't be as helpful when you are zeroing small buffers of non-round sizes


That's what (b) is about.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: