Sender: port-i386-owner@netbsd.org

> > BTW, what kind of temperatures are people seeing on the CPU's?  My
> > 900 MHz Athlon runs around 60=B0C, around 70=B0C with full load (accord=
ing
> > to the mb sensor).  The lowest reading (somewhere on the mb, I think)
> > is 25=B0C.  Should I be worried?
>=20
> silicon chips can SURVIVE 130-140 degres and work at 90-100 degrees
> properly. it's about in-core temperature, on packing-box temperature
> - that measured - would be 20-30 degrees lower due to heat resistance.
>=20
> so 60-70 degrees is maximum acceptable. for short - if you can't keep your
> hand on chip more than few seconds it's too hot

All this discussion brings back memories from my computer/electrical
engineering classes for my major.  The two rules of thumb to remember
here are:

- For most components, current roughly doubles for every 10deg C rise
  in temperature. (I would have to dig out one of my textbooks to find
  the exact formula, but it is an exponential).

- For every 10deg C rise in temperature, you figure on halving the
  lifetime of a component.

Things get really complicated when you consider that some, if not many
manufacturers determine the clock speed of a chip on a batch-by-batch
basis through testing.  However, since I prefer to spend my money on
new "toys" rather than fixing ones I already have, I tend to be
conservative and not over-clock.