Timers and timing, was: MySQL Performance 6.0rc1

[Massimo Lusetti]

On Thu, 2005-10-27 at 21:40 -0400, Mike Jakubik wrote:
> On Thu, October 27, 2005 8:54 pm, ender wrote:
> 
> > CPU: AMD Athlon(tm) MP 2800+ (2133.42-MHz 686-class CPU)
> > FreeBSD 6.0-RC1
> > SMP kernel,
> > #cpu            I486_CPU
> > #cpu            I586_CPU
> >
> >
> >
> > null function: 0.00471
> > getpid(): 0.18755
> > time(): 1.18135
> > gettimeofday(): 1.17575
> 
> Thats pretty odd, considering the following:
> 
> FreeBSD 7.0-CURRENT #0: Sat Sep 24 21:12:10 EDT 2005
> CPU: AMD Athlon(tm) Processor (1410.21-MHz 686-class CPU)
> 
>        null function: 0.00145
>             getpid(): 0.26670
>               time(): 0.85759
>       gettimeofday(): 0.83653

Yep, and on 5.3-STABLE of 12/2004 on a Xeon 2.8G

       null function: 0.00460
            getpid(): 0.41424
              time(): 0.55854
      gettimeofday(): 0.54748


But, while repeating, I've seen values of time() and gettimeofday()
changing even to more then 2 secs.


For the records here is the result on a Linux kernel 2.6.12-1.1380_FC3
(Fedora Core 3) on a P4 1.8G

       null function: 0.00947
            getpid(): 0.00988
              time(): 3.80196
      gettimeofday(): 4.04731


-- 
Massimo.run();