This is the mail archive of the
gcc-help@gcc.gnu.org
mailing list for the GCC project.
Re: Performance problem
- From: Christoph Bartoschek <bartoschek at gmx dot de>
- To: gcc-help at gcc dot gnu dot org
- Date: Sun, 5 Oct 2008 11:07:18 +0200
- Subject: Re: Performance problem
- References: <c55009e70809211420l7f649d8ah9e027c55b9221567@mail.gmail.com> <48D733AD.3000603@cppm.in2p3.fr> <c55009e70809220137n4f2fd1ddrb254b9e6479ffaca@mail.gmail.com>
Am Montag 22 September 2008 schrieb Åukasz Lew:
> My code uses also rusage - second most accurate source of timing I
> found on linux.
> It the same as used by time command line tool.
I use always wall time measured with gettimeofday on a machine that is
reserved for my benchmark. My experience is that CPU-time measurement on
linux can vary by 100% on machines that have high load. Even for machines
with no load the CPU-time measurement varies by about 20%.
Christoph