Discussion

Discussion on Article:

Started by: Spoonbender | Date 10/24/04
Comments: 4 | Last Comment:  10/25/04

[1-4]

1. Damn, 67W? If they keep this up, they'll be able to give the Pentium M a run for its money... :D
[Posted by: Spoonbender | Date: 10/24/04]

2. LOL...The Pentium-M at 2Ghz does 21W in typical thermal dissipation...
Its less than 30W under max thermal dissipation. Intel will never reveal the actual number, but you can tell the max has to be below 30W because that's a notebook requirement. (excluding "desktop replacement" class portables). That's also using the 90nm process.

All P6 based CPUs have been cooler than AMD's K7 and K8 lines by quite a bit. (usually half the power dissipation)...Compare the old PIII and Athlon lines clock for clock in power dissipation. Athlon is always hotter while being slightly faster.

But it does show one thing...The Pentium 4 is a VERY inefficient design. It thrives on more speed at the cost of everything else. Its like a tank with a turbine engine. Sure its powerful, but look how much fuel it consumes.

So in the end, its not just about the manufacturing process, its also about the processor architecture.

Pentium-M 90nm = good, less than 30W at most. (but expensive)
A64/etc 90nm = OK
P4 90nm = BAD! BBQ! Anyone for Satay Prescott? :)

Personally, Intel should just phase the P4 out ASAP. You can't push it any further than 4Ghz, why not use "Yonah" (dual-core P-M) for 2005 release of dual-cores from Intel? Why have Smithfield? (dual-core P4)

How much different are the chipsets? They've already shown they can install a Pentium-M class CPU into P4-class chipset, so why not?
[Posted by: 22 | Date: 10/24/04]

3. First off, this is being said about chips that are still very low in clockspeed. When we got the A64 to 2.6ghz we saw things rise to 104. And that mind you is with designers being told to prepare for around 105, so personally I think that's not enough safety margin. And furthermore if you believe they will come out with high speed dual cores that fit in that 105W envelope, yoiu're crazy :) Of course I've been wrong before! But I don't think they'll manage dual cores going at anything like 2.2-2.4 and still keep that under 105W. That means all those nice NF4 mobos you have now, would technically be compatible with dual cores, but only the lower speed varieties. Anyone who wants dual core at higher speed is looking at mobo upgrade as well. That's how I see things panning out in 2005.

$.02
[Posted by: Anemone | Date: 10/24/04]

4. Is the 67W for the 2.2 GHz 3500+? If so, that'll put the 3000+ around 55W. If the wattage is compared to a 2.2 GHz Barton, which wattage is the right one for the Barton, 76.8W or 60.4W?
Also, is the Cool 'n' Quiet from the CG-stepping onwards also running normal 130nm Athlon64 processors at 1.0 GHz with a IIRC lower power usage than the older 800 MHz versions?

From some of the articles I've read which take the 90nm processor's power consumption into consideration, I've come away impressed with the numbers they show. 67W seems like a lot, but the 89W of the 130nm sounds very high as well.
[Posted by: Esben | Date: 10/25/04]

[1-4]

Add your Comment

Add your Comment

Name/Nickname
Your Comments