Will Overclocked Northwoods die young?

[MarkAtTheZone]

I've seen one post of a moderately overclocked Northwood that didn't last long even at a rather safe 1.625 volts (this is what I run mine at!). Aceshardware has suggested that the more closely spaced conductors in the Northwoods might cause overclocked chips to fail early due to electromigration shorts. It would be a great marketing strategy to make wonderfully overclockable chips that always died young. Is Intel laughing their asses off at us in expectation of selling us new chips in a few months? ....Mark

 

[KenAF]

Mark,

Well, Intel has a no-questons-asked return policy under their three year warranty, so it doesn't really matter whether you fry your processor, now does it? Do you really intend to keep any processor you buy today for more than three years?

That said, I must have read posts from 300 different people with the Northwood, a high percentage runniing 2.4GHz to 2.7GHz, and that's the first report of failure that I have seen. Most likely, the processor was defective in some way.

 

[BigJ]

Overclocking shortens the life of any processor, so this is nothing new. What is suprising is that TheThugs processor died so quickly. It could have been a defective chip to begin with, cause many people have been running for a few days longer and at 1.65volts then him with no death of their processor. I think the best thing to do in this situation would be to buy retail, so even IF the processor does die, you have a 3 year warranty covering you. If what Aceshardware is true, it really does pay to spend the extra $ to get a retail instead of OEM processor.

 

[mschell]

The question is are you really "overclocking" when you run a 1.6 at 2.2GHz. It's entirely possible that Intel yields are good enough that all Northwoods can do 2.2GHz, they are just marked down to fill market demands. AMD chips are the same - I'm running a XP+1500 at the +1900 clock level, default 1.75v, again they just marked down the chip for cheapskakes like me who can't afford a +1900.

 

[THUGSROOK]

actually right now we are overclocking because there is no P4 133/533 spec yet.
his 2.2ghz is not produced the same way as a retail 2.2ghz. actually hes faster then a retail 2.2 in most things cause of his bus speeds.

in the p3 100/133fsb days - youd have been 100% correct tho

 

[Duvie]

mschell, that IMO is exactly right....

The only disadvantage of getting such a slow processor ie like a p4 1.6a eventhough you know the whole line (1.6a to 2.2a) can get to around 2.6ghz with very little cooling is that unless you have the best motherboard many can run into huge problems with too large of a fsb overclock to achieve it...I think we are seeing it now...The p4 being locked only leaves fsb tweaking and many boards like asus c version have usb device issues above a certain fsb, restart issues on gigabyte board, ram unless offered many divider features either runs underclocked or way out of spec...


1.6 is a great deal but 150fsb to get to 2.4ghz may be pushing it for many unless they use a pci locked board like abit, but according to review sites likely couldn't reach that stabily anyways with the abit board...2.6ghz takes a 163 fsb and most likely a pci lock and pc2700ddr would be necessary...


I am shooting for a 2.4ghz from my 1.8a....I want the 533fsb and try to attain it with a voltage of 1.625 or less...Based on 2.2ghz reviews a 2.4ghz should take an athon xp 2200+ or 2300+....The 2200+ can likely b attained on a 2000+, but take more expensive cooling, power, and generate a lot more heat.

The 1.8a retail box ran me 208.00...An athlon 2000+ xp retail box (the cheapest I could find on pricewatch) was 282.00

 

[resinboy]

I re-joined the Dark Side, and just finished a 1.8A, and an Asus P4B266 install. I had to increase core voltage to 1.65 to get stable at 133 FSB (2.4 gig). Seems solid after a bunch of tests, but I am also worrying about the extra voltage. ALSO, even after updating to the latest bios ( 1005), I still have a "frozen" temp sensor readout ( mb bios, as well as using Asus' software).
Since these new core steppings haven't been around all that long, should I step the voltage back down for a while to be safe till we see how this goes?
ALSO , as far as their warranty policy, how will the use of Artic silver, instead of the thermal pad effect it??( you swap boards, you have to scrape the original thermal pad anyways)???

Resinboy

 

[JellyBaby]

No one really knows except Intel engineers and these guys are always locked up in dark rooms. My guess is that if moderate over-volting has a chance of seriously damaging the processor during its expected life, Intel would have taken steps to prevent this. Perhaps through warnings, perhaps through stricter standards, etc.

Overclocking is always a risk but until P4s start dying en-mass I wouldn't worry too much.

 

[KenAF]

resin,

The extra voltage doesn't matter at such levels...its heat that matters, and even at 1.8 volts, the P4 still produces less heat than the Athlon XP 2000+. People aren't having failures at 1.8 to 1.9 volts at 2.8GHz to 3.1GHz, so I don't see why anyone should worry at 2.4GHz, 2.5GHz, or even 2.7GHz. Now, if you were to crank it up to 2.1 to 2.3 volts, as some have done for 3.6+GHz overclocks, then I might get worried...