On 12 May 2006 07:32:58 -0700
"llothar" <llothar@web.de> wrote:

> If i have a huge CPU bound task like a large compile job where
> utilization is near 100%.
> How long do the fans need to go to full speed ?

Never. I've never seen the fans go beyond 40%, even with both CPUs
going at full blast for an hour. Ambient temperature is 24ºC, the
machine is a Blade 2000 with two 900MHz Cu CPUs compiling Firefox
1.5.0.3 (from prtdiag -v):

---------------------------------------------
Location Sensor Status Speed
---------------------------------------------
+stem-fan-slot system-fan okay 31%
+/cpu-fan-slot cpu-fan okay 34%
+r-supply-slot power-supply okay 100%

Temperature sensors:
-----------------------------------------------------------------------------
Location Sensor Temperature Lo LoWarn HiWarn Hi Status
-----------------------------------------------------------------------------
+em-board/cpu0 Die 78C -10C 0C 88C 90C okay
+em-board/cpu0 Ambient 26C -10C 0C 40C 60C okay
+em-board/cpu1 Die 72C -10C 0C 88C 90C okay
+em-board/cpu1 Ambient 24C -10C 0C 40C 60C okay

> My iMac G5 is going from okay to extremely noise just after a few
> seconds 100% CPU useage. This means during each compile and this
> on-off is much more worse then a constant mid level noise. I can't
> tell you how much i hate this machine.

I agree - I had a laptop that did exactly the same thing and it was far
more disrupting than a constant noise. I was very pleased when I could
replace it with an Acer Ferrari 4000 which has a constant fan.

My Blade 2000 is far quieter than a cheap Athlon 2000+ based Linux PC
that shares my desk, even with 2 10,000 rpm FC-AL disk drives.

--
Stefaan A Eeckels
--
"Object-oriented programming is an exceptionally bad idea which
could only have originated in California." --Edsger Dijkstra