Prescott to hit 120 Watts
 

Prescott to hit 120 Watts

http://www.overclockers.com/tips00493/

Bad news for air coolers, good news for watercoolers!

I'd have to say it's not good news for anyone except AMD.

Even with watercooling or phase change cooling those procs are going to build heat really fast when pushed.

I've seen speculation these chips may go as high as 175 watts if pushed real hard.

Now a couple of those in a house will cut down on the winter fuel bill.

This is the most effective marketing campaign for me that Intel has had in ages. No one else likes a challenge? I like big Watts!

LOL

Actually it is good news for all of us IMO. Now HS manufactures will have to pull double time in R&D. If nothing else this should raise the level of HS tech yet again. 175 watts though? Overclockers are going to have to find higher power pelts! :)

I wonder how Zaleman will take this news with that monster heat piped case they can't seem to get to work for over 6 months now. Nothing like a little extra challenge as pH pointed out. :D

change the color ?
blue is cooler

haha that's funny on a couple of levels, Bill

LOL, if I see one in blue I will laugh my ass off!

The real problem here is going to be..
How do you get almost 200Amps
into the part without melting some pins or trace. At 200A, even just a little resistance is going to get hot, fast! A poor connection is going to go incandescent as it heats up and the resistance increases, causing more heat...

90nm parts are going to sub 1V Vcore's,
so every Watt is more than an Amp of current.
Designers are going to have to think about cooling the Vcore supply, and the power connections.

This reminds me of something I read in one of Drexler's nanotech book about future super computer CPU cores being a 3d chip, maybe a foot on a side, two sides have copper cables, carrying as much as a megawatt of power, one side is a high pressure water inlet, the other side, live steam outlet 1 side for fiber-optic IO and one side to screw it into a support structure.
We're getting there..

That's why the majority of pins on CPUs are power & ground. And, that's why Intel is abandoning sockets for Land Grid Arrays. No pin resistance, no socket resistance, fewer solder joints, just matching contact pads on motherboard and chip carrier.

After that, live heatsinks? Signal lines at the bottom of the chip carrier; on the top, power and ground. The voltage regulator built into and cooled by the heatsink.

Fun stuff to contemplate.