Does Optimization Effect BogoMips Value?

Does Optimization Effect BogoMips Value?

[Justin Piszcz]

Three identical Virtual Machines with three different compiler flags to
compile the entire distribution, in this case, Gentoo 2004.1.

However, why are the Bogomips values different (up to +82 points
different)?

Linux tux 2.6.7-gentoo-r6 #1 Wed Jun 30 05:56:57 EDT 2004 i686 Intel(R)
Pentium(R) 4 CPU 2.60GHz GenuineIntel GNU/Linux
Gentoo 2004.1 Compiled With: "-pipe -mcpu=i386": 1515MB FREE
Calibrating delay loop... 4866.04 BogoMIPS

Linux tux 2.6.7-gentoo-r6 #1 Wed Jun 30 14:01:09 Local time zone must be
set--see zic manu i686 Intel(R) Pentium(R) 4 CPU 2.60GHz GenuineIntel
GNU/Linux
Gentoo 2004.1 Compiled With: "-Os -pipe -mcpu=i686": 1673MB FREE
Calibrating delay loop... 4882.43 BogoMIPS


Linux tux 2.6.7-gentoo-r6 #1 Wed Jun 30 13:58:50 Local time zone must be
set--see zic manu i686 Intel(R) Pentium(R) 4 CPU 2.60GHz GenuineIntel
GNU/Linux
Gentoo 2004.1 Compiled With: "-O3 -pipe -mcpu=pentium4 -march=i686 -msse
-msse2 -fomit-frame-pointer -funroll-loops -ffast-math": 1635MB FREE
Calibrating delay loop... 4800.51 BogoMIPS

Re: Does Optimization Effect BogoMips Value?

[Rik van Riel]

On Wed, 7 Jul 2004, Justin Piszcz wrote:

> Three identical Virtual Machines with three different compiler flags to
> compile the entire distribution, in this case, Gentoo 2004.1.
> 
> However, why are the Bogomips values different (up to +82 points
> different)?

It really doesn't matter for performance. A higher
bogomips value just means the system will spin more
delay loops to wait the same amount of time...

For more details, see http://funroll-loops.org/

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan

Re: Does Optimization Effect BogoMips Value?

[Nagendra Singh Tomar]

On Wed, 7 Jul 2004, Rik van Riel wrote:

> On Wed, 7 Jul 2004, Justin Piszcz wrote:
> 
> > Three identical Virtual Machines with three different compiler flags to
> > compile the entire distribution, in this case, Gentoo 2004.1.
> > 
> > However, why are the Bogomips values different (up to +82 points
> > different)?
> 
> It really doesn't matter for performance. A higher
> bogomips value just means the system will spin more
> delay loops to wait the same amount of time...


If it can spin more delay loops in the same time, then probably it can do 
more useful work also in the same time.


> 
> For more details, see http://funroll-loops.org/
> 
> 

-- 



-- You have moved the mouse. Windows must be restarted for the 
   changes to take effect.

Re: Does Optimization Effect BogoMips Value?

[Horst von Brand]

Justin Piszcz <jpiszcz@lucidpixels.com> said:
> Three identical Virtual Machines with three different compiler flags to
> compile the entire distribution, in this case, Gentoo 2004.1.
> 
> However, why are the Bogomips values different (up to +82 points
> different)?

Because the CPUs run at slightly different speeds? Because the cristals
that give PCs wall clock are inaccurate (82 in 4866 is 1.6%)?
-- 
Dr. Horst H. von Brand                   User #22616 counter.li.org
Departamento de Informatica                     Fono: +56 32 654431
Universidad Tecnica Federico Santa Maria              +56 32 654239
Casilla 110-V, Valparaiso, Chile                Fax:  +56 32 797513