[Clave]

All I know is that my back room gets pretty warm after a while - Apple need a slap if they put that PSU in there for no good reason

[StratComm]

Originally posted by Clave
But even so, I just dont get it - as far as I know the IBM processors actual use less power than a P4....

Have you actually seen the heat sink on the G5?  Those chips run dangerously close to cooking most of the time, which is a pretty good indication that they actually use a lot of power.  They don't clock as fast, but there's way more going on inside them than a P4.

[Descenterace]

My recent upgrade to an Athlon64 3500 has led to headaches re: heat. The CPU gets very hot even when idle, and this is after I bought a ThermalRight XP-90 heatsink, a 30CFM 92mm fan and some Arctic Silver 5. And lapped both heatsink and CPU heatspreader...

Hence my planned upgrade to watercooling and a ground-up rebuild of my system. I just hope I can get a decent price for my current rig...

[Clave]

Like the G5? http://www.apple.com/powermac/design.html

It's just a bit scary when stuff is running so damn hot, there must be another way...

BTW: my office comp (G4) is running at 57.1C right now (ambient 25.2C)

[Nix]

Originally posted by Descenterace
My recent upgrade to an Athlon64 3500 has led to headaches re: heat. The CPU gets very hot even when idle, and this is after I bought a ThermalRight XP-90 heatsink, a 30CFM 92mm fan and some Arctic Silver 5. And lapped both heatsink and CPU heatspreader...

Hence my planned upgrade to watercooling and a ground-up rebuild of my system. I just hope I can get a decent price for my current rig...

How hot exactly?  I'm running a similar setup, a 3000+ 64 and a Venus 12 on my system, and I idle at 38C, if my room is at actual room temperature.  It'll top out at 52C, maxed out running the UD agent.  

And you lapped the heatspreader on the processor?  I wouldnt think that they'd need that sort of treatment.  Heh. I've never actually lapped my heatsinks at all and I get great performance.  
Another thing to consider too, how are you putting the  AS5 compound on?  If you slather the entire chip with it, it might not be transferring the heat as efficiently (though, when you think about it, it probably doesnt matter).  Arctic Silver's website has been updated with a new way to apply the compound.  Not only does it make perfect sense, it makes it way easier when you have to take the heatsink off.  One guy I knew slathered it on thick all over the chip.  Couldnt get it off unless we maxed out the CPU for a bit, got the temps up, shut down the system, and got the chip off.

[Descenterace]

I've run into that same problem; nearly took the pins off the chip the first time I removed it, and that was just with silicon grease!

I applied the AS5 in the 'correct' way: a very small blob in the middle of the heatspreader, then lower the XP-90 on to it (insofar as it's possible to lower a heatsink that can really only be clipped down if it pivots around one edge...). It makes a nice thin circle of compound over the core.
My temps are about the same, but I bought the XP-90 to allow me to overclock this system. Seems to have been a waste of money.
On the other hand, the fan doesn't seem to be doing much. I can actually remove the CPU fan entirely and get no difference in temperature.
Given that I'm using an ABIT (Actual is Below Indicated Temperature) mainboard, I'm not inclined to trust the built-in temperature monitor.
This system's performance should change when that massive Delta fan gets here in a week's time... 140CFM and 57dB of turbine.

[Nix]

Dood, if you're gonna do that, make a variable resistor switch, turn it down when you dont need it topped out all the time.  
I dont trust automatic thermal controls.

[Descenterace]

I'm investing in a 20W fan controller for this purpose. I would build my own, but I'm lazy

[phatosealpha]

Are you experiencing any actual stability problems though?  I ask cause I'm running something very similar w/ an xp-90 or a MSI neo-2 platinum, and the temperatures readings are just plain wrong.  If your CPU is a winnie, the problem is probably just very bad readouts from the the thermistor.  Something changed between the newcastles and the winchesters, apparently it causes quite a few motherboards to just give wildly innaccurate results.

[Descenterace]

It's a Newcastle. And I know it becomes unstable at readings around 65-70, which is about right for Athlon64s.