Good news, Acehardware did some research on some Intel documents regarding the Prescott's heat production and it seems that the previous estimates of 100watt were to high.
The production CPU's will run cooler than the enginering samples!
Here's a part of the article:
"Based on previous Intel documentation, 68A is slightly more than the maximum sustained current that was observed using a large suite of applications. In other words, it is highly unlikely that any useful realworld application will ever exceed 68A sustained. 78A is the maximum current that can occur but its not sustained. Therefore, depending on the voltage of Prescott, we have a thermal design (real world maximum) power of ~82W, with a maximum power of ~93W for 1.3V, or 75.5W TDP and 86W max for a core voltage of 1.2V. A far cry from the 103W rumors that still exist today. Also, since this is the design document for S478 Prescotts, it could be expected that this is the thermal envelope for the entire range of S478 Prescotts, for 3.4GHz or maybe even 3.6GHz."
You can find the rest at: http://www.aceshardware.com/
Prescott to produce 'only' 82watt, not 103!
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
-
- SPCR Reviewer
- Posts: 8636
- Joined: Sat Nov 23, 2002 6:33 am
- Location: Sunny SoCal
Considering that Intel's at least 2 months behind their original release date, it's pretty easy to figure that they've been burning the midnite oil to try and reduce the voltage requirements for this CPU.
82W is what the P4 3.0 and 3.2 NW's put out so current cooling technology should be able to handle Prescott with ease, assuming this 82W estimate is in the ballpark.
82W is what the P4 3.0 and 3.2 NW's put out so current cooling technology should be able to handle Prescott with ease, assuming this 82W estimate is in the ballpark.
-
- Posts: 1283
- Joined: Wed Sep 03, 2003 1:35 am
- Location: Sweden, Linkoping
Sigh! This does worry me a bit.it is highly unlikely that any useful realworld application will ever exceed 68A sustained.
This is like testing with Folding@home and say "Nope, we could just reach 80% of TDP, so lets set this value as TDP".
If you did like that you would not be able to handle apps that push to the max like CPUBurn or Prime95. Many processors today put out a lot less than the rated TDP, but I don't see this as a good reason to lower the TDP.
The use of TDP is to give a value that can be used when designing cooling systems (e.g. heatsinks etc) so you don't have to redo lots of work when they raise the clock 0.2GHz.