[OT] Decibels and the ear [was: Re: [plug] Hot and bothered CPU hankers for cool breeze]

[Derek Fountain]

> The "rule of thumb" (that Cameron provided) implies that
> increasing the intensity 400-fold would cause a 2.6-fold increase in
> loudness. Derek's experience would seem to indicate that 400 machines
> were not 400 times louder than one machine.

Obviously in a lab that size a lot of the machines were some distance away 
from any one point - someone standing in the middle of the room might be 
barely able to hear a machine in the furthest corner, even when the room is 
otherwise silent. And since the room also contained several thousand high 
performance disks spining at full speed, not to mention the extraordinary air 
conditioning required to keep everything cool, getting a precise handle on 
how loud the servers were probably isn't possible. The theories and 
algorthims put forward in this discussion were interesting, but for this 
example they probably get washed out by the, er, noise. :o)

Actually, it's discussions like this one which sometimes make me wish I'd 
studied physics or chemistry or maths rather than CS. I wish I understood 
more!

-- 
> eatapple
core dump