Degradation: Would you buy a second hand wolfdale ?

[sablabra]

It seems fairly risky running 1.53 V through a 45 nm CPU for 24/7 use. We really don't know to much about these early batches yet, but I have no doubt in my mind that such high voltage could damage/degrade the chip.

Thinking back, I have always been cautious with voltage as the die shrimps.
My old athlon xp 2600+ would easily run at 1,70 V or more.
My old x2 3800+ ran happily at 1.55 V for 24/7 use (roughly 15 % increase in Vcore). Not very much for a 90 nm chip.
My current Q6600 is running at 1,46 V for 24/7 use (roughly 12 % increase in Vcore given 1,3 V is stock). I wouldn't go much higher for 24/7 use.

This is all done with water cooling. With the new 45 nm CPUs I probably wouldn't go higher than 1.35-1.40 V. Looking over my quick calculations above, a 15 % increase in Vcore would mean 1,38 V, which seems about right to me. However, please keep pumping volts 1,5+ volts through your CPUs, I'm eager to see if more CPUs die

 

[Monkeynut]

The lower process means less volts, and people are pumping the same volts through the 45nm chips as they did on the Q6600s. Then they cry when their CPU dies because of silly volts.

 

[hernaldo]

The lower process means less volts, and people are pumping the same volts through the 45nm chips as they did on the Q6600s. Then they cry when their CPU dies because of silly volts.

Pretty much the case, seems they're almost forgetting the fact

 

[Foehammer2003]

Well i have spent best part of the evening scouring google, and dont think I have ever read so many posts in one night before.

My motivation? Just purchasing a E8400 from MM atm, my conclusion... the muppets killing (cooking? ) their chips are putting waaayy too much voltage through, and even this is in the minority. I guess i shall find out soon enough, but im happy that as long as i stay under 1.4 or slightly less then I would be fine, hell I might even run it at stock

Encouraging to see nearly all clocks @ 4Ghz minimum on air though, most I found was some guy got 5.4!!!

I'm not greedy though with my OC's (lol sig testifies to that!) just a bit to give me my something for nothing. And will my chip have been cooked before I purchase? well apparently and hopefully not, and dont we all run the same risk regardless of model for whichever CPU people buy 2nd hand?

 

[mattyrigby00]

i think it might be an issue of that with 65nm etc you run into heat problems before the volts are too high, but with 45nm the volts can be set to dangerous levels without it overheating, but still being damaged.

 

[Devious]

And that there isnt a program giving correct temps for wolfidales.

Ive seen enough people putting 1.7v through their G0 chips on air which, pro rata must be at or above the levels are putting thorugh these wolfdales but i canot remember seeing one degrading chip post.

 

[Tooks]

And that there isnt a program giving correct temps for wolfidales.

Ive seen enough people putting 1.7v through their G0 chips on air which, pro rata must be at or above the levels are putting thorugh these wolfdales but i canot remember seeing one degrading chip post.

What went before regarding voltages on 65nm chips is completely irrelevant.

CPU voltages don't scale like that.

As others have said, it's a 45nm process, with a max voltage printed on the side of the box at 1.225v. The VID on my old one was 1.125v if I remember, so 1.5v/1.6v is pushing it too far in my opinion. Mine did 3.8Ghz at 1.25v, and I was reluctant to try any more voltage to get it higher.

Googling does reveal a few posts about 'degrading' wolfdales, true, but then we should be grateful that other people have got there before us and wrecked their chips with high voltages, so we can learn their lessons.

I think we've all become a little complacent about how well Intels 65nm Core2's and Quads clock and how they seem impervious to voltage, so maybe people should be a little more cautious now and learn the best way to clock the 45nm parts instead of assuming they can bang through whatever voltage they used to on the previous generation?

 

[Hesky82]

I dunno, I got fed up of wimping out with low voltages and went right for 1.55v and my e8400 responded very well to it and has been rock solid stable ever since, tested prime small fft and it still happily passes with no errors. temps never go over 56C under load on my water.

I think its easily to get confused and think your system was stable at a said lower voltage and clock only to realise you never really made the attempt to make sure it was actually 100% stable before you went romping on for max clock

One thing that I have found is that 500+ FSB really plays havok with my system and can at times cause my system to not be game stable, yet it will run ANY cpu stress test no problem. Basically I feel its my mb letting me down not this cpu.

 

[james.miller]

i think its safe to assume that you shouldn't risk anything over a ~125% overvolt on the cpu.

 

[Big.Wayne]

my e8400

Haha nice one Hesky, you swapped from a monster overclocked quad core to a monster overclocked dual-core?

 

[Hesky82]

Haha nice one Hesky, you swapped from a monster overclocked quad core to a monster overclocked dual-core?

Still have both, just in an annoying catch 22. Can’t use both so must sell one.

Should I pop the quad back in for future proofing or stay with the dual and have a stupid fast dual core which I know I can fully use 100% of the time right at this time.

My Q6600 is a hot cookie at 4GHz, but I just have this annoying feeling it will come good when the apps I use actually work out their multi threading, but as of this minute I simply can’t get 100% use out of it, having just 1 or 2 software suites that actually fully utilise it just doesn’t warrant keeping it over the dual.

 

[Big.Wayne]

My Q6600 is a hot cookie at 4GHz

It also probaby draws enough wattage from the mains under load to keep a small nation lit!

Sell the Quad now as a golden-sample before the Yorkfields tear up the place!

 

[Hesky82]

With my dual running more voltage I bet it’s not that far behind! my machine still lights up the 4th Power indicator on my ups/mains conditioner, so the system is still drawing pretty much the same power as it did with the quad at 1.4v ish.

Pushing them to the limit really does kind of ruin all the eco friendly vices these new 45nm have, but when they respond so well to high vcore it just seems rude not to, low temps are also comforting.

 

[chronic86]

i've been running my e8400 @ 4ghz with 1.35(1.325 after vdroop) since it was released. and my pc has never been switched off for more then 1 day since i've had ! its rock solid at the moment but i do not plan on taking it any further then my current overclock. i'm gunna run orthos again to make sure its still stable. not had a sniff of a problem so far tho!

fingers crossed

 

[prosser13]

One of the big differences between 45nm and 65nm is that with 45nm temperatures have been all but removed from the equation - you can pump stupid amounts of volts through chips and even cheapo watercooling will keep up fine...

Because of this, people are upping the voltage without thinking about the consequences - the default VCore for a E8400 is around 1.2V, and people are talking about running them at 1.5V?! That's a 30% voltage increase, and I wouldn't be surprised if the fact that it's a smaller manufacturing process means that even a small increase in voltages speeds up degradation a lot more than in 65nm =/

The old 65nm 'safe voltages' are still being used for 45nm, even with the stock voltages being much lower. Anyone putting 1.5V through a E8400, especially with air cooling, is asking for trouble.