You are not logged in.

#1 2006-06-02 04:42:28

uberGeek
Member
From: Near Chicago
Registered: 2004-01-07
Posts: 65

Crazy bogomips value

I just noticed this, but last time I looked at /proc/cpuinfo, both CPU's had a reasonable value (around 5600 - 5700) now I have 5637 on cpu0 and 157,087 on cpu1!

Wondering if anyone else has seen this?

I suspect this happened after updating to the current kernel, 2.6.16.19, which only differs the stock kernel in that 64GB is on, as opposed to 4GB.  I have the previous kernel, and I might check it this weekend.  Everything else seems to be working fine on a very stuffed system, so I'm not so much worried as just curious.


carb $ now; uname -a; uptime
Sun 03Jul11 14:54:25 CDT -0500
Linux carb 2.6.17-ARCH #1 SMP PREEMPT Fri Jul 7 09:15:53 CEST 2006 i686 Pentium III (Coppermine) GenuineIntel GNU/Linux
14:54:25 up 1463 days,  9:24,  1 user,  load average: 1.00, 1.00, 1.00

Offline

#2 2006-06-02 08:48:33

brain0
Developer
From: Aachen - Germany
Registered: 2005-01-03
Posts: 1,382

Re: Crazy bogomips value

uberGeek wrote:

I suspect this happened after updating to the current kernel, 2.6.16.19, which only differs the stock kernel in that 64GB is on, as opposed to 4GB.

2.6.16.19 has 4GB support, not 64GB.

Offline

#3 2006-06-02 11:54:21

JGC
Developer
Registered: 2003-12-03
Posts: 1,664

Re: Crazy bogomips value

You've got a P4 HT CPU? I get different values on "CPU0" and "CPU1" also for the bogomips. Please take note: these things are bogus tongue

Seems HT screws up as a fake CPU again, nothing to do about it. Having less bogomips isn't really important anyways.

Offline

#4 2006-06-02 13:41:33

uberGeek
Member
From: Near Chicago
Registered: 2004-01-07
Posts: 65

Re: Crazy bogomips value

I built my own kernel, based on 2.6.16.19 from ABS, with the only change being a few edits to the config, to turn 64GB on and 4GB off.  I've done this for the last 4 or 5 kernel updates.  Thus the statement: "...only differs the stock kernel in..."

I have a pair of Opterons.  If by bogus you mean they are not a "benchmark" , then I agree.  My 2.8GHz Opterons fold protiens a lot faster than my 3GHz P4's.  I am under the impression that the kernel uses the bogomips for some internal timing, so I'm wondering if this could cause problems.  Again, I've not experience any problems yet, well, that I know of anyway smile


carb $ now; uname -a; uptime
Sun 03Jul11 14:54:25 CDT -0500
Linux carb 2.6.17-ARCH #1 SMP PREEMPT Fri Jul 7 09:15:53 CEST 2006 i686 Pentium III (Coppermine) GenuineIntel GNU/Linux
14:54:25 up 1463 days,  9:24,  1 user,  load average: 1.00, 1.00, 1.00

Offline

Board footer

Powered by FluxBB