ppc: bogomips at 73 when CPU is at 1GHz

ppc: bogomips at 73 when CPU is at 1GHz

[Junichi Uekawa]

Hi,

I've noticed the very log value on bogomips on self-compiled 2.6.16.16
on iBook G4.

[06:12:59]ibookg4:~> cat /proc/cpuinfo
processor       : 0
cpu             : 7447A, altivec supported
clock           : 1066.666000MHz
revision        : 0.1 (pvr 8003 0101)
bogomips        : 73.47
timebase        : 18432000
machine         : PowerBook6,5
motherboard     : PowerBook6,5 MacRISC3 Power Macintosh
detected as     : 287 (iBook G4)
pmac flags      : 0000001b
L2 cache        : 512K unified
pmac-generation : NewWorld


It was somewhat higher on 2.6.14.


Apparently I'm not the only person who noticed something similar; 
but I can't really read spanish:

https://listas.hispalinux.es/pipermail/linux-ppc-es/2006-May/000820.html


Am I missing something or is everyone seeing this?


regards,
	junichi

Re: ppc: bogomips at 73 when CPU is at 1GHz

[Benjamin Herrenschmidt]

On Wed, 2006-05-17 at 06:18 +0900, Junichi Uekawa wrote:
> Hi,
> 
> I've noticed the very log value on bogomips on self-compiled 2.6.16.16
> on iBook G4.

 .../...

Normal. We don't use CPU bogo loops for short timings, we use the CPU
timebase instead, thus you now get bogomips derived from the timebae
frequency of the machine.

Ben.
 

Re: ppc: bogomips at 73 when CPU is at 1GHz

[Junichi Uekawa]

Hi,

> > I've noticed the very log value on bogomips on self-compiled 2.6.16.16
> > on iBook G4.
> 
>  .../...
> 
> Normal. We don't use CPU bogo loops for short timings, we use the CPU
> timebase instead, thus you now get bogomips derived from the timebae
> frequency of the machine.

Thanks.

I was tracking down a possibly unrelated problem with USB wlan device
(GW-US11H) and suspected this to be a cause of the problem.


Incidentally, if anyone has any success using USB wlan device
(GW-US11H), I'd like to hear that ... it's rather inconvenient to be
at Debian Conference with a wireless connection that hangs your
machine after 5 minutes.

http://www.netfort.gr.jp/~dancer/diary/daily/2006-May-17.html.en#2006-May-17-05:18:32


regards,
	junichi

Re: ppc: bogomips at 73 when CPU is at 1GHz

[Gabriel Paubert]

On Wed, May 17, 2006 at 06:18:19AM +0900, Junichi Uekawa wrote:
> Hi,
> 
> I've noticed the very log value on bogomips on self-compiled 2.6.16.16
> on iBook G4.
> 
> [06:12:59]ibookg4:~> cat /proc/cpuinfo
> processor       : 0
> cpu             : 7447A, altivec supported
> clock           : 1066.666000MHz
> revision        : 0.1 (pvr 8003 0101)
> bogomips        : 73.47
> timebase        : 18432000
> machine         : PowerBook6,5
> motherboard     : PowerBook6,5 MacRISC3 Power Macintosh
> detected as     : 287 (iBook G4)
> pmac flags      : 0000001b
> L2 cache        : 512K unified
> pmac-generation : NewWorld
> 
> 
> It was somewhat higher on 2.6.14.
> 
> 
> Apparently I'm not the only person who noticed something similar; 
> but I can't really read spanish:

Well I can (I'm not a from Spain, but been living here
since 1986).

> 
> https://listas.hispalinux.es/pipermail/linux-ppc-es/2006-May/000820.html
> 
> 
> Am I missing something or is everyone seeing this?

Don't worry, bogomips are bogus. Now on PPC they are related to
the processor timebase frequency, which on G3 and G4 processor
is the bus frequency divided by 4.

	Regards,
	Gabriel

Re: ppc: bogomips at 73 when CPU is at 1GHz

[Wolfgang Pfeiffer]

On Wed, May 17, 2006 at 06:18:19AM +0900, Junichi Uekawa wrote:
> Hi,
> 
> I've noticed the very log value on bogomips on self-compiled 2.6.16.16
> on iBook G4.
> 
> [06:12:59]ibookg4:~> cat /proc/cpuinfo
> processor       : 0
> cpu             : 7447A, altivec supported
> clock           : 1066.666000MHz
> revision        : 0.1 (pvr 8003 0101)
> bogomips        : 73.47
> timebase        : 18432000
> machine         : PowerBook6,5
> motherboard     : PowerBook6,5 MacRISC3 Power Macintosh
> detected as     : 287 (iBook G4)
> pmac flags      : 0000001b
> L2 cache        : 512K unified
> pmac-generation : NewWorld
> 
> 
> It was somewhat higher on 2.6.14.

Nothing to be worried about, as it seems:

http://lists.debian.org/debian-powerpc/2006/03/msg00085.html
http://lists.debian.org/debian-powerpc/2006/04/msg00126.html


HTH

Best Regards
Wolfgang

-- 
Wolfgang Pfeiffer: /ICQ: 286585973/ + + +  /AIM: crashinglinux/
http://profiles.yahoo.com/wolfgangpfeiffer

Key ID: E3037113
http://keyserver.mine.nu/pks/lookup?search=0xE3037113&fingerprint=on

Re: ppc: bogomips at 73 when CPU is at 1GHz

[Sven Luther]

On Wed, May 17, 2006 at 06:18:19AM +0900, Junichi Uekawa wrote:
> Hi,
> 
> I've noticed the very log value on bogomips on self-compiled 2.6.16.16
> on iBook G4.
> 
> [06:12:59]ibookg4:~> cat /proc/cpuinfo
> processor       : 0
> cpu             : 7447A, altivec supported
> clock           : 1066.666000MHz
> revision        : 0.1 (pvr 8003 0101)
> bogomips        : 73.47
> timebase        : 18432000
> machine         : PowerBook6,5
> motherboard     : PowerBook6,5 MacRISC3 Power Macintosh
> detected as     : 287 (iBook G4)
> pmac flags      : 0000001b
> L2 cache        : 512K unified
> pmac-generation : NewWorld
> 
> 
> It was somewhat higher on 2.6.14.
> 
> 
> Apparently I'm not the only person who noticed something similar; 
> but I can't really read spanish:
> 
> https://listas.hispalinux.es/pipermail/linux-ppc-es/2006-May/000820.html
> 
> Am I missing something or is everyone seeing this?

2.6.16 doesn't use the some kind of do-nothing loop anymore to set the
bogomips value, but some internal cpu counter or something such. I don't know
the details.

As such, the bogomips value changed radically, and what you see is normal.
bogomips values are bogus anyway, so this change should not have any other
major effect.

Friendly,

Sven Luther