[prev in list] [next in list] [prev in thread] [next in thread]
List: linux-smp
Subject: Re: Bogomips and MMX was (Re: RedHat 5.0 + Tomcat + Intel 166Mhz MMX)
From: "Manfred Brands" <m.brands () fugro-intersite ! nl>
Date: 1997-12-19 16:28:13
[Download RAW message or body]
I lost about 50 BogoMips, by changing an option in the same kernel.
It all depends on alignment of the 'bogo-mips' loop.
Whether it is aligned on a 'cache' line or not.
There is a patch at Big Mama (http://www.huwig.de/linux/mama)
which aligns the loop at 16 bytes to give a consistent number for the
same processor.
The difference between MMX and non-MMX is Intel's bigger level-1 cache
and jump prediction.
Manfred.
----------
> From: Kimmo Laine <kl2398@turkuamk.fi>
> To: Chris Pirih <proverbs@wolfenet.com>
> Cc: linux-smp@vger.rutgers.edu
> Subject: Re: Bogomips and MMX was (Re: RedHat 5.0 + Tomcat + Intel 166Mhz MMX)
> Date: Friday, 19 December, 1997 11:53
>
>
> Earlier I send those result ... that was with kernel 2.0.32 .. I
> compiled new one from 2.0.33 source and here are those results
>
> >cat /proc/cpuinfo
> processor : 0
> cpu : 586
> model : 4
> vendor_id : GenuineIntel
> stepping : 3
> fdiv_bug : no
> hlt_bug : no
> f00f_bug : yes
> fpu : yes
> fpu_exception : yes
> cpuid : yes
> wp : yes
> flags : fpu vme de pse tsc msr mce cx8 apic mmx
> bogomips : 330.96
>
> processor : 1
> cpu : 586
> model : 4
> vendor_id : GenuineIntel
> stepping : 3
> fdiv_bug : no
> hlt_bug : no
> f00f_bug : yes
> fpu : yes
> fpu_exception : yes
> cpuid : yes
> wp : yes
> flags : fpu vme de pse tsc msr mce cx8 apic mmx
> bogomips : 330.96
>
>
> It has came something like 100 more bogomips. How the hell to analyze
> these figures ? and what REALLY has happened ?
>
>
> **** Kimmo Laine ****
> **** kimmo.laine@turkuamk.fi ****
>
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic