prescott temps a good thing?????

I know that this thread might belong in the INTEL section but I would rather have a few AMD types opinion on the matter. After recently reading a review of the new FX-53 compared to the prescott and the mobile barton I couldnt help being blown away by this :

"After solving the temperature problems, I managed to get it (prescott) to sit loaded at a rather modest 77'C"

Earlier in the review he talked about the prescott having a load temp of 88'C. I had heard a few things about the prescott running hot but 88'C under load is incredible. Imagine the posibilities..........

as you turn on your computer in the morning you slide over to the built in frying pan and quickly fry some eggs up for breakfast (frying pan of course would be conected directly to the heatsink).

a cold winter day.... "Dad, can you please open up the computer case its feezing outside and the house is a bit chilly" 10 seconds later "Ohhhh thats much better"

Microsoft and Intel have always talked about having computerized appliances in the near future...... the future is closer that you think.

In all seriousness, how can a chip even function at 88'C? I thought blue screens started appearing at around 65'C. WOW intel really screwed up this one.

can you post the link to that review?
Thanks,
Aaron

towards the end he make the statement about getting it down to 77C at idle. Does anyone have any experiance with these chips??? DO they really get that hot?
http://bit-tech.net/review/309

In all seriousness, how can a chip even function at 88'C? I thought blue screens started appearing at around 65'C. WOW intel really screwed up this one.

Click to expand...

You hit the nail on the head. CPUs really shouldnt run over 65c* under stressed conditions. Over these temps the silicon begins to litererly tear itself apart through a process called "electromigration".

It will run at those temps just fine. It a question of how long... At best maybe a year, maybe 2. However eventhough IT can run at those temps, other parts cannot. Namely the mobo and videocard.

Several Intel guys have had their mobo destroyed by Prescott. The guys over at [H]ard|OCP actaully got the mobo to weld itself to the test bed.

From an AMD fanboy perspective, Im very amused. Pride has killed Intel. They thought they were too good for SOI. Now its bitten them in the ***. Now the only pplz in the world that can create SOI are now joined at the hip to AMD. (IBM / AMD)

Intel is flat out screwed until they radically re-design their processors. They WILL be forced to take AMD's logic. (quality > quantity)

Its glad to see that AMD has the opportunity to enter a new golden age!

_______________________

towards the end he make the statement about getting it down to 77C at idle. Does anyone have any experiance with these chips??? DO they really get that hot?

Click to expand...

From what Ive heard, Prescotts run a full 10-20c* higher then their NW / CW counterparts. 77c* IMHO is very believable

Skip all the way to the "IN CONCLUSION" part of the article. What is intel doing about this??

Hmmm..... I wonder if Prescott looks so bad because its thermal throttling during benchmarking. If it is, I can't imagine the temps if it was running at full clock speed all of the time.

I agree with Sentential; Intel needs an architectural change to ramp with AMD in the near future. Lucky for them, they already have this architecture - I believe Intel will begin to market Dothan processors to desktop customers. In fact, changing their nomenclature to get away from MHz makes it easier to convince customers that a 2GHz Dothan can be faster than a 3.4GHz P4.

Why redesign your architecture when you can borrow (ala P4EE) from a different market segment to stay competitive?

I was at CeBIT, and got a chance to see an LGA775 system in demonstration with a prescott CPU. Man, that thing was hot under load. One could feel the heat two feet from the stand.

There was a Sudhian review in which they tried using a prescott in a shuttle... bad idea. The PSU which had run at 50C before with a 3.2C rampaged on to 90C before popping with the prescott.

As has been stated before, the CPU can handle high temperatures, just not for more than 1-2 years. The more immediate problem, however, is not only the electrical, but also thermal stress placed on the surrounding components. SEVERE stress.

In all honesty, intel would have been better trying to push the northwood to 4. It wouldn't have run NEARLY as hot as a prescott and would have kept AMD in check.

I don't know what their plans are... as of now, they seem to be releasing northwoods and prescotts at equal speed grades. Whether that continues at 3.6 and 3.8 will be directly tied to the outcome of the 90nm process shrink revisions.

Leakage current really does tax regulators.

The prescott is using stretched silicon which is a germainium hybrid and are designed to be able to run at higher temps. The prescott at its most concentrated Ge is around 30% Ge to 70% Si.

I am not the expert but most IC spec do indicate the max operating ambient temp. is 85 degree. So if the chip itself is 78 degree it should be fine.

I just post something in the IC datasheet, a very normal audio IC:
Specification, ambient: 0~75 degree
Operation, ambient: -25~+85 degree,
Absolute Maximum ratings: Junction, +200 degree

So I guess the chip will really melt down and damage itself permanentely when it is something like 100 or more.

**repeat: i am not expert, just IMHO**

This is w/ +15C air and Zalman7000CNPS CU running at full speed on AI7 mobo


[email protected] 1.6V

I also know that according to specs most amd chips have a max operating temps of 85c (my old 1800+) but we all know that anything over 60c and (not in every chip) is where errors start happening. In my testing to get a quieter case I turned down my fans to a minimum (temps shot up) and found that to be prime95 stable over 60c for more than a few minutes was very hard. I found this to be true at a number of voltages and speeds. Soooo, if the max temps of both chips were the same you would think that they would behave similarly as they approached their max operating temps. I would like to know if anyone can prime95 torture test their prescott at these high temps (60c +) or if errors will occur as well.

Sorry for the ramble.

long time no see macci

Ok so it was designed to operate at higher temperatures than standard - but what happens to the mobo components over a sustained period!

macci, I am very interesting to know your configs, because I am considering to get a 2.8E.

btw do you mean this chip is 80C when ambient temp is only +15C? wow if that's true, the chip may burn when summer comes (we got +35C room temp in summer)

over 60c and (not in every chip) is where errors start happening

Click to expand...

Really depends on how the mobo reads the CPU temp. 60C on ASUS P4C800 equals to like +80C on IC7. Certain AMD boards use external sensors while others read the internal temps etc etc.
But in any case the 'Scottie is one hot chip

btw do you mean this chip is 80C when ambient temp is only +15C?

Click to expand...

Yup and its almost +80C in BIOS - no idea how hot it would have got w/ CPUBURN or similar software..
Ambient temp +15Cish (open windows mod ).
That particular 2.8E would run 3550MHz superpi 1m stable at default volt aircooled and 3781MHz at 1.55V aircooled. Useing vaporphase (Mach2 R404) and 1.55V it would pass tests at 4.35-4.44GHz. Used P4C800-E and InWin 300W PSU for those vaporphase clocks.

OC Detective said:

Ok so it was designed to operate at higher temperatures than standard - but what happens to the mobo components over a sustained period!

Click to expand...

I think the component most sensitive to the temp is the E-Cap. Normally it is rated +85C and higher grade rates +105C. But the life shorts badly when temp increases. I saw some spec said if work under rated remp, the life of the E-cap would be around 2000hrs only(life time=when the capacity degrades to half of original spec).

I also eagerly to know if Prescott will work properly under high temp like 80C....I am planning to buy one.

prescott runs hot but it can handle the heat.

Prescott may handle the heat (I'm pretty sure it'd have to with temps like that ), but I don't think many other components can.... My bet's that BTX has a little more than just a monster HS to save just the CPU


JigPu

yeah imagine a btx cooled prescott with no fan on it. impossible sez i.