• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Coolers for Intel Prescott Begin to Emerge; Ready to Dissipate 103W Now

Typical power consumption of PGA478 Prescott processors will be over 100W, while for the higher-speed parts in LGA775 package the figure is heading towards roughly 120W.:Q
 
At this rate vapor phase change cooling will be mandatory in a few years ..
rolleye.gif
 
LOL Rectaflier...BUy 2 and you would only be at 90w and still less!!!😉


What is my current 2.6c@3.25ghz at now with 1.63v actua vcore???

 
I thought we ended that orb cooler fad. It dosen't work well so why does Intel want it? Of course their orb cooler shows some promise. But I'll bet with that fancy aluminum shape it would be hard to fabricate.
 
Originally posted by: Duvie
LOL Rectaflier...BUy 2 and you would only be at 90w and still less!!!😉


What is my current 2.6c@3.25ghz at now with 1.63v actua vcore???
Click to expand...
I wonder about that too. I'm @ 3.34 GHz @ 1.57 actual Vcore. Its not hard to air cool at all. Nearly silent too.
 
Dissipation numbers.

CPU.............................Core..............Typical .............Max
Pentium 4 3.06 GHz ...1.55 V..............81 W...........+/- 105 W

One other very important point to remember when comparing the .13 micron p4's to the .09 micron prescotts is Die size, and I can only assume the prescotts will have a smaller die (I have heard rumored around 80 mm2, compared to the .13 P4 which is around 146 mm2, 237mm2 for the p4EE) , which will exponentially increase the amount of cooling needed for the cpu (assuming they output similar heat). Someone else here (mikewarrior i think is his name, or i am sure others) could probabaly give us a more detailed explanation of the exact figures involved in cooling the smaller cores.
 
Originally posted by: klah
I wonder if IBM/AMD would license SOI to them...
Click to expand...

The currently assumed cause of Prescott's high power consumption is due to leakage. The partially-depleted SOI implementation used by AMD/IBM do very little in solving this problem. As leakage becomes a significant concern at .09um, AMD will likely suffer from leakage issues as well. Since no clock/voltage values were assigned to the 45W, 0.09um Opterons, it is too early to assume that AMD will not have hot chips too.
 
Originally posted by: Accord99
Originally posted by: klah
I wonder if IBM/AMD would license SOI to them...
Click to expand...

The currently assumed cause of Prescott's high power consumption is due to leakage. The partially-depleted SOI implementation used by AMD/IBM do very little in solving this problem. As leakage becomes a significant concern at .09um, AMD will likely suffer from leakage issues as well. Since no clock/voltage values were assigned to the 45W, 0.09um Opterons, it is too early to assume that AMD will not have hot chips too.
Click to expand...


I forgot where I saw the news, but AMD said that the Athlon 64 FX running at 2.4Ghz will dissapate 45watts. I can try and find the link when I get home from work.
 
Originally posted by: Rectalfier
Originally posted by: Accord99
Originally posted by: klah
I wonder if IBM/AMD would license SOI to them...
Click to expand...

The currently assumed cause of Prescott's high power consumption is due to leakage. The partially-depleted SOI implementation used by AMD/IBM do very little in solving this problem. As leakage becomes a significant concern at .09um, AMD will likely suffer from leakage issues as well. Since no clock/voltage values were assigned to the 45W, 0.09um Opterons, it is too early to assume that AMD will not have hot chips too.
Click to expand...


I forgot where I saw the news, but AMD said that the Athlon 64 FX running at 2.4Ghz will dissapate 45watts. I can try and find the link when I get home from work.
Click to expand...

The whole point of SOI is to prevent leakage - an oxide layer isolates the substrate and the transistors, which dramatically reduces leakage. I don't see why SOI wouldn't significantly help AMDs transition to .09um.
 
Originally posted by: sellmen
The whole point of SOI is to prevent leakage - an oxide layer isolates the substrate and the transistors, which dramatically reduces leakage. I don't see why SOI wouldn't significantly help AMDs transition to .09um.
Click to expand...

It does not prevent gate leakage, which is the dominant factor starting at 0.09um. This is why both AMD and Intel, and everybody else, are working hard on high-K dielectrics, fully depleted SOI and multigate transistors. Unfotunately these won't be available for awhile so gate leakage is going to affect everybody.
 
Originally posted by: sellmen
Originally posted by: Rectalfier
Originally posted by: Accord99
Originally posted by: klah
I wonder if IBM/AMD would license SOI to them...
Click to expand...

The currently assumed cause of Prescott's high power consumption is due to leakage. The partially-depleted SOI implementation used by AMD/IBM do very little in solving this problem. As leakage becomes a significant concern at .09um, AMD will likely suffer from leakage issues as well. Since no clock/voltage values were assigned to the 45W, 0.09um Opterons, it is too early to assume that AMD will not have hot chips too.
Click to expand...


I forgot where I saw the news, but AMD said that the Athlon 64 FX running at 2.4Ghz will dissapate 45watts. I can try and find the link when I get home from work.
Click to expand...

The whole point of SOI is to prevent leakage - an oxide layer isolates the substrate and the transistors, which dramatically reduces leakage. I don't see why SOI wouldn't significantly help AMDs transition to .09um.
Click to expand...
Keep in mind that AMD's SOI is only the "partially-depleted" technique; the full benefits come from fully-depleted, so they aren't nessisarily going to see significant boosts at this point.
 
Originally posted by: Accord99
Originally posted by: sellmen
The whole point of SOI is to prevent leakage - an oxide layer isolates the substrate and the transistors, which dramatically reduces leakage. I don't see why SOI wouldn't significantly help AMDs transition to .09um.
Click to expand...

It does not prevent gate leakage, which is the dominant factor starting at 0.09um. This is why both AMD and Intel, and everybody else, are working hard on high-K dielectrics, fully depleted SOI and multigate transistors. Unfotunately these won't be available for awhile so gate leakage is going to affect everybody.
Click to expand...

Accord to
This link, SOI is the reason that AMD is ahead of Intel in the .09 transition, and why they may be shipping earlier than planned.
 
Originally posted by: Accord99
Originally posted by: sellmen
The whole point of SOI is to prevent leakage - an oxide layer isolates the substrate and the transistors, which dramatically reduces leakage. I don't see why SOI wouldn't significantly help AMDs transition to .09um.
Click to expand...

It does not prevent gate leakage, which is the dominant factor starting at 0.09um. This is why both AMD and Intel, and everybody else, are working hard on high-K dielectrics, fully depleted SOI and multigate transistors. Unfotunately these won't be available for awhile so gate leakage is going to affect everybody.
Click to expand...

I remember reading about new Intel tri-gate technology in the process in the CPU mag a few months ago. It was an interesting read although I didn't understand a lot of it. The graphs/pictures helped 🙂 I heard IBM and AMD are also working on multigate redesigns, although Intel is the only one working on a tri-gate design.
 
Originally posted by: Markfw900
Accord to
This link, SOI is the reason that AMD is ahead of Intel in the .09 transition, and why they may be shipping earlier than planned.
Click to expand...

It's talking about transistor leakage... but at 0.09um, gate leakage has a much greater impact on power consumption. And AMD is not ahead of Intel, Intel will ship sometime in the first quarter of 2004, AMD will ship 6-9 months after.
 
I don't give a sh!t how much heat that thing puts out! If it's the fastest chip out, I'll take it. Now if two chips performed the same and one ran cooler, of course I'd take teh cooler running chip. 😉 But I think the new Prescotts will have alot of overclocking headroom. 🙂
 
I found the link to Athlon64 2.4Ghz producing 45Watts. They also state that AMD's SOI is key to keeping the transistor leakage down.

The story is on The Register, you can find the link here
 
Originally posted by: JackBurton
I don't give a sh!t how much heat that thing puts out! If it's the fastest chip out, I'll take it. Now if two chips performed the same and one ran cooler, of course I'd take teh cooler running chip. 😉 But I think the new Prescotts will have alot of overclocking headroom. 🙂
Click to expand...

I think that the Prescott's will be something like the first Thoroughbred. With the delay and the reports of heat problems, I don't think there will be much overclocking headroom. Intel will surely overcome this hurdle with a Prescot B, but not before they finish feeding us their original Prescotts.

With a 45Watt disapation, and reports of being ahead of schedule, the 90nm Athlon64 will be much more attractive, and may prove to be the next "Northwood".

 
You must log in or register to reply here.

TRENDING THREADS

Back
Top