Computer power consumption

[Walt_Crawford at notes.rlg.org]

I think this is relevant to Web4Lib, since the Web won't do much for you if
fallacious power-consumption claims mean that you have to shut down your
machines...

http://www.wired.com/news/technology/0,1282,41623,00.html

The gist of this story: a widely-publicized 1999 study of current and
anticipated power consumption by computers and related technologies (which
forecast that such devices would account for HALF of U.S. power consumption
a little ways down the road) was just plain wrong.

For example: the study assumed that a typical PC, including monitor,
consumes 1,000 watts.
Reality: a typical Pentium-III with 17" CRT uses about 135 watts (most of
that being the display).

(And, of course, power consumption changes. My old LaserJet at home used
100+ watts idling, more than twice that when printing. My newer one uses
3-5 watts idling, 60 or so printing--and most people are migrating to
inkjets, which use at most a few watts.)

(This part isn't in the Wired News article, but was in a Marketplace
discussion yesterday afternoon: the study assumed a server-farm
per-square-foot consumption that's at least four times as high as actual
_measured_ per-square-foot power consumption at server farms. Oh, and by
the way, the researcher has ties to the coal industry...)

The 1999 study's sponsor, the Greening Earth Society, asserts that "CO2 is
GOOD for us," and is directly linked to fossilfuels.org, an association to
propound the benefits of coal etc...

Not to say that power conservation isn't important--but phony numbers don't
help.