Archived from groups: alt.comp.periphs.mainboard.abit (
More info?)
It would be impossible to say with any certainty that a 1MHz reduction is
down to electron migration in the CPU....Any of a host of things could cause
this tiny change, for instance; slight loss of thermal paste between
heatsink and CPU, dust on the heatsink, dust on the fan, change in
performance of components on the motherboard or in the psu (eg capacitors),
change in room temp, updated bios etc etc....
This is also assuming over the 6 months no hardware has changed which for
most tweakers/modders/overclockers would never be the case
--
*****Replace 'NOSPAM' with 'btinternet' in the reply address*****
"- HAL9000" <gumpy@mail.org> wrote in message
news:3bop60lloo6njmlv158rla2ls6m3nshpbj@4ax.com...
> Of course the electron migration will depend greatly on the chip at
> hand. And, critical to the test is the precision of your measurement.
>
> I really don't know if the migration is linear with voltage or
> quadratic with voltage. A couple of google searches, others seem to
> indicate that it is a quadratic function. That is, increase the
> voltage a little and the amount of electron migration will increase a
> lot.
>
> My experience is that overclocking my Duron 1G to 1.85 showed about a
> 1 MHz reduction after 6 months. This is why I suggested the method.
>
> Also, I used to run an old super socket 7 motherboard and ran the cpu
> at 2.4 volts for overclocking. The cpu was supposed to be run at 2
> volts as I recall. Over the 3 years that I used it as my regular
> machine, the upper end stability limit very noticeable deteriorated.
> Every year or so I had to go into the bios and lower the fsb slightly
> to regain stability. Towards the end, the cpu wouldn't even do its
> rated freq without overvolting slightly. But then again, the cpu
> never really overclocked well.
>
> Forrest
>
> Motherboard Help By HAL web site:
> http://home.comcast.net/~hal-9000/
>
>
> On 01 Apr 2004 05:12:19 GMT, private1964@aol.com (PRIVATE1964) wrote:
>
> >>I would guess that after 6 moths the max fsb freq will be 1 or perhaps
> >>2 MHz lower.
> >>
> >>Forrest
> >
> >I was overclocking a 1900+ Palomino at a 1.9 vcore for around a year and
I
> >didn't see any loss of performance. It's default vcore was 1.75.
>
>