Yes, I suppose in a non-nit way, your statement is correct or close
enough for approximation. Just didn't seem right from a basic physics
POV. Another thing occured to me besides heat and light. Radio waves!
Computers do create enough EM interference that you can hear it on a
cell or portable phone (especially 2.4GHz frequency sets). Most of that
is probably the power supply, but I'm pretty confident the CPU also
radiates radio frequencies.
To introduce some specific terminology, to the extent that the CPU is
surrounded by by a solid, all of the energy that is dissipated by the
CPU has to escape either as a signal with useful and energy-carrying
information going out through a pin, be converted into a phonon or
escape as a photon.
There must be whole classes of quantum states that correspond to the
signals that make it onto PC board traces, and I truly have no idea
what to call them, since some of them must correspond to correlated
pairs of electrons, one on each trace if differential signalling is
being used, and they would qualify to be called a "particle" in the
same way that the Cooper pairs of superconductivity qualify to be
called particles. If it can be imagined, surely someone has written a
paper about it.
Electromagenetic radiation at any frequency is carried by photons.
The phonons are what we normally think of as heat. Very little energy
gets out of the CPU in any fashion whatsoever without at some point
being scattered and at least temporarily converted to a phonon. If
you _really_ wanted to look at all the possibiliteis, Feynman diagrams
would be helpful for categorizing all the exotic combinations of
energy pathways (taxonomy: ugh!).
There is, I am sure, a finite quantum transition amplitude for
production of every conceivable quantum particle that anybody has ever
heard of, but on balance it is clear that they contribute nothing that
anybody knows how to calculate to the energy budget of the CPU.
If you've ever once actually played with Feynman diagrams, which are
almost necessary even to keep track of all the terms in the equations,
you will quickly discover that there is both more and less to all of
this than meets the eye.
Nature knows nothing (and I believe that this statement would have
stood up to scrutiny by Prof. Feynman himself) of the discrete
interactions that field theoreticians and solid state physicists draw
on blackboards.
Instead, there are an infinite number of interactions, involving an
infinite number of Feynman diagrams, going on all of the time. Some
of those interactions involve evanescent intermediate states that are
as real as the xerox toner on the printed page, but they violate
either energy or momentum conservation or both so we call them
"virtual" particles. All of this can happen without a CPU present.
In fact, all of it can happen without anything present at all. Go
google for vacuum fluctuations if you don't believe me.
Prof. Feynman might give me an argument about the following, but much
of what physicists routinely peddle as identifiable physicial
phenomena (an example being virtual photons) are really nothing more
than artifacts of the way he chose to go about solving the
(non-linear) equations of QED. This is clear enough to me because
I've used the same formal apparatus in a classical context, where it
is clear that you can write the equations down, you can draw the
Feynman diagrams, and you can draw useful conclusions, but no new
physics spring into existence because you choose to solve non-linear
equations in a particular way.
Your posts reverberate with the confusions of someone who has been
bamboozled in a physics classroom. *Nobody, nobody, nobody* knows
enough physics to answer your question with absolute, final certainty.
If the "standard model" is to be believed, the most ordinary F=ma type
calculation of freshman physics involves a Higgs field that is carried
by bosons that no one has managed to capture in a measurement. That
way lies madness. Be happy if you are in computer science or
electrical engineering and don't have to sit through arguments about
these things.
It is probably a significant number, but I doubt it is more than a
fraction of the CPU itself (1's of watts vs 10's of watts?). Otherwise
there'd be lots more chatter about having the right battery for your CPU
so that voltage stepping losses don't eat into your battery life budget.
It is, in any case, more important than any kind of exotic particle
interaction you and I both understand less well.
RM
