We came to this conclusion several posts ago :)

It felt irrelevant to me what timecounter h/w / driver I was using -
as stated repeated times, after reporting about this I was -just
curious- on what changed from 4.1 to 4.2, and that has been perfectly
elaborated already by Otto Moerbeek and yourself. What's interesting,
though, is that the P3/600 and the Sempron/1.9 machines - again, both
running 4.2-stable/i386 - both use the i8254 h/w + driver:

kern.timecounter.tick=1
kern.timecounter.timestepwarnings=0
kern.timecounter.hardware=i8254
kern.timecounter.choice=i8254(0) dummy(-1000000)

For your pleasure, dmesg for the Sempron-machine is here:
http://pastebin.com/m4fce6b00

What you want to make of this is up to you. I am -perfectly- satisfied
with the explanation that was given earlier in the thread.

The C2D machines running in the park at work are out of my reach for
many days ahead, as is the P3/600, so I can't please you with any
dmesg's of those for now.

-SD

On Sat, Mar 15, 2008 at 11:10 AM, Stuart Henderson <stu@spacehopper.org> wrote:
quoted text
> On 2008-03-15, Jonathan Thornburg <J.Thornburg@soton.ac.uk> wrote:
>  > Apart from these nits... my results on a Thinkpad T41p (i386 Pentium M)
>  > running 4.2-stable are (test program compiled with gcc 4.2.0, -g -O2):
>  > ... with 'apm -H' in effect (clock speed 1.7GHz): 2.92 seconds
>  > ... with 'apm -L' in effect (clock speed 0.6GHz): 3.98 seconds
>
>  I think this proves that the hardware and/or driver for the
>  timecounters most people are testing this with are quicker to read
>  than whatever the original poster is using (but doesn't want to
>  tell us what it is - I don't understand why someone goes to the
>  trouble of providing a test program but won't include the
>  always-requested dmesg, but there you go).
>
>  There are (if I counted correctly) 11 possibilities for timecounter
>  on i386 (though not all in the same machine :-) - note that the
>  kern.timecounter.hardware sysctl is /not/ read-only ...