[Shootout-list] fannkuch
Bengt Kleberg
bengt.kleberg@ericsson.com
Wed, 25 May 2005 08:18:04 +0200
On 2005-05-24 19:36, John Skaller wrote:
...deleted
> On x86 I am getting +/- 1 second for the same 8 second programs.
if i am not mistaken that is +/- 12.5%
as a theoretical exercise i would like to show how i am thinking about
using the timer granularity for shortest acceptable run time:
if the run time is 0.0051 it will be measured as 0.01
if the run time is 0.0149 it will be measured as 0.01
we have +/- 0.0049 seconds for the same 0.01 second programs.
if i am not mistaken that is +/- 49%.
bengt