Memorize a Dead GEM QX9650 Retail | 5.65G @ DI

[VictorWang]

weekend is coming, so I decided to play ag.
but after I take out the cpu and install on my mobo, comes only SAD

cpu can boot at any clk at it used to be.
but can't even show the boot menu with all default settings.

changed 3*HDDs and 3*mobos...all the same
but change another cpu, everything is working good.
though i dont wanna accept the reality, but it is that my GEM cpu dead

might be L2_Cache's prob, 'coz when I test Spi_Dos, it said: not converge in sql.

I was about to break 6Ghz with this chip and I'm pretty sure about that.
but....... the only thing I can do is to show some WaterCooling and DI results


Part I: Watercooling , ambient=17c, swiftech aprogee gt, d5, 3*12 copper hx with 3*12cm 1700rpm fan.



4.95G boot from bios:






in OS now and get ready for benching:




let's pull up clk:






3Dmark06:





Part II, DI section: Pot temp= -50~-57c.








max cpu clk under DI:



Part III: LN2 test.... never have change to do that 'coz i'm making new pots these days.
but right after my pots finished, CPU dead!
think this must be the most sad thing in the world for me.
i dont care about money, but such a good cpu dead just break my heard

[Zeus]

That's a painfull loss man.

Any idea what caused it's dead?

[eva2000]

Ouch Victor, sorry to hear

When you had cpu on sub zero, did you use dielectric grease in socket / back side of cpu ?? Maybe too much dried or old dielectric grease on back side of cpu ? When i tried my old Q6600 B3 on phase on commando at first didn't boot up but fans were spinning etc..

Problem was too much dielectric grease LOL... cleaned it all off and booted with no probs..

Tried CPU PLL volts on AUTO ??? used many volts on CPU PLL with this cpu ??

I hope it's something simple and resolvable and not dead.

Maybe it needs a rest and will come back to life after a time out hehe

Good luck

[VictorWang]

Ouch Victor, sorry to hear

When you had cpu on sub zero, did you use dielectric grease in socket / back side of cpu ?? Maybe too much dried or old dielectric grease on back side of cpu ? When i tried my old Q6600 B3 on phase on commando at first didn't boot up but fans were spinning etc..

Problem was too much dielectric grease LOL... cleaned it all off and booted with no probs..

Tried CPU PLL volts on AUTO ??? used many volts on CPU PLL with this cpu ??

I hope it's something simple and resolvable and not dead.

Maybe it needs a rest and will come back to life after a time out hehe

Good luck

i just finished these, so u can see how much i'm expecting on the chip.



never used >1.65v vPLL ,
max vcore used = 1.8v under DI.
last benching sesion = 4.8G 3Dmark06 under watercooling.
after that, the chip rested for a whole week and laying on my shelf together with Asus ME.


but now...everything is ready but the cpu dead

[eva2000]

Darn it then, such a gem gone.. R.I.P.

I think this is the first or second only reported dead QX9650 on XS ?

[Leeghoofd]

Sorry to hear this, can anyone share a light on why this has happened... never heard much of dead CPU's on this forum...

Where's the link to the other one sir Eva2000 ?

[VictorWang]

Darn it then, such a gem gone.. R.I.P.

I think this is the first or second only reported dead QX9650 on XS ?

far as i know, there are 4~5 9650 dead. and 1*9770 dead.

[xgman]

No one wants to admit that 1.8v can kill this thing? Let me be the first then. It would seem to me that this is no mystery. No matter what the cooling, anything over 1.7v is probably a quick way to either kill or at minimum shorten the life of this cpu.

[Zeus]

Stoolman killed one @ 2.175V and i know that Coolaler already killed 2 45nm cpu's.

[VictorWang]

shamino killed his QX9770 > 2v / 6.08G

[Leeghoofd]

So what do you guys think is safe ? 1.6 max ? I'm using watercooling but apparently these cpu's can be very misleading to add more voltage for more Mhz as they don't produce as much heat as the other quads...

[cowie]

No one wants to admit that 1.8v can kill this thing? Let me be the first then. It would seem to me that this is no mystery. No matter what the cooling, anything over 1.7v is probably a quick way to either kill or at minimum shorten the life of this cpu.

sure death runs at -80 and close to 2.v will lead to faster failure thats a no brainer .
but it sucks when it happens i think you just try not to think about it happening.

[Kunaak]

the lower the nm the lower the safe voltage is, isnt it?

thats the way its been in the past with the move from 130 to 90, and 90 to 65...

[Aivas47a]

the lower the nm the lower the safe voltage is, isnt it?

thats the way its been in the past with the move from 130 to 90, and 90 to 65...

That makes a lot of sense. I never worried too much benching kentsfield at 1.6 -1.65v with water but I think I'll be keeping it under 1.6 for yorkfield. Unless I get rich.

[Falkentyne]

People didn't want to admit that 'GNDS' (gradual northwood death syndrome) affected Core 2 Duo's too. Heck, some people (i'm not giving names) didn't even want to admit it affected NORTHWOODS. That BS cost me hundreds of dollars, buying chips, listening to IDIOTS here (no offense to the people who are smart) telling me that 1.65+v is perfectly safe on air if you have good cooling, and watching my chips degrade in OC from 3.9 to 3.8... Other people saying "1.75v or less on air is safe"... O_O and that northwood that started at 3.9 wouldnt even post anymore at 3.7...

Of course when I was PROVEN right, by OTHER People posting, where were the haters?

The same thing happened with core 2's too. My X6800 lost a few FSB (or needed another .25v) after just running it at 3.6g @1.50 and 3.7g@ 1.575v for awhile. Yet people were too stubborn to admit that that problem still happens.

This did not happen to this extent with the old Pentium 3 coppermines, unless you REALLY went above the spec. And I never saw any posts about P3s degrading like that. But to put things in perspective, I saw degradation on a northwood at JUST 1.575v, which is a voltage that EVERYONE says is 100% safe. VERY very minor and its taken many months, but it is there.

So you do this at your own risk. I don't know what voltage is really 'safe', but it obviously FAR below Intel's absolute maximum.

Back to topic: Sorry for your loss, Victor Youre probably right that the cache went south.....

[dinos22]

extremely sorry to hear that Vic


this is not really proof falkentyne
it's just a dead chip and we don't know the cause

others were all high vcore related from stoolman, shamino, coolaler

[Timbosan]

how unfortunate, right when everything else is ready too. I hope its resurrectable somehow.

[Falkentyne]

extremely sorry to hear that Vic


this is not really proof falkentyne
it's just a dead chip and we don't know the cause

others were all high vcore related from stoolman, shamino, coolaler

I'm not talking about Victor's chip.
I'm talking about raising voltage in general.

Check the forums and do a search, and check ocforums, too.
I'm not the only one who has seen degradation in an overclock on a c2D. There was a post LAST WEEK about someone who saw their max overclock decrease at about 1.6v on a core 2 duo. (the post is on ocforums--go find it).

I saw it on 4 northwood P4's, and my C2D X6800. What more proof do you need, honestly? Me wasting thousands of dollars on chips and then me inviting you to my house so you can watch the overclock degrade at 1.55v?

Sure, if you pay for it.

It's been PROVEN that going too high on voltage can degrade an overclock. It's called electromigration and this is PROVEN. What is NOT proven is at what voltage it will occur, and this seems to vary from chip to chip. But I had this happen at 1.575v on a northwood, when max operating voltage was 1.55v. This did NOT happen on P3's for such a voltage increase; it took a LOT more % on those.

My problem is people saying "keep it at less than 1.65v on air and you'll be fine", like a "go fever" type of mentality--the same thing that doomed Challenger in 1985. Instead of a "YMMV: it was fine for me but I can't guarantee all chips will function reliably at x.y voltage", people think that just because a problem doesn't happen to them, it doesn't exist. I can go on and on, giving many more real life examples of this mentality, but there's no point. People will always believe what they want to believe. Just because I'm not [H] or extreme like some of you guys breaking world records does NOT make my statements any less valid. Again these chips are far less 'sustainable' (varies per sample) to high voltage increases than older chips. I remember putting 3.3v into a Pentium MMX (3.3v was the default for an old pentium, IIRC), to overclock from 166 to 262.5, and I never saw the chip degrade through its lifespan. I had more problems getting that board running an 83 mhz FSB...and not having things corrupt. Remember THOSE days?

[Plaicd]

I have heard intel will be using something called high-k metal instead of silicon for new chips(not sure if is in use yet).
Any idea if that will make cpu's die faster when overvolted or be able to take more voltage or no difference?

Sorry to see you lost that cpu

[Zeus]

Falkentyne, you're talking about degrading of a chip, i'm sure nobdy in here will deny that it happens, we see the same thing happening to RAM.

Victor's chip died just out of the blue, i think it's a different case.

It looks like thes 45nm are more prone to die on high voltage than 65nm.

[Dumo]

Sorry to hear that VW. Once in a while It happens to all of us....Shorted the power switch on mobo and fans all work for a second then all quiet.

Time to hunt for another 5.8 ghz 9650

[VictorWang]

Sorry to hear that VW. Once in a while It happens to all of us....Shorted the power switch on mobo and fans all work for a second then all quiet.

Time to hunt for another 5.8 ghz 9650

no good chip, no energy to benching more.
it's been the first time in the last 3years that I slept in my bed enjoyed a 20hours dream

anyway, it's the first Intel Chip I killed since I start playing PC 14years ago.
and with no clue how it can dead in this way

chip is dead but my overclking life will still go on.
time to call Intel for a RMA now

[LaGUNaMAN]

... chip is dead but my overclking life will still go on.
time to call Intel for a RMA now

So that's how the Pro's do it. j/k

[Carsten]

"time to call Intel for a RMA now " - you must sure hope that they don't read this thread,

----------CB

[Pete]

Victor,

You've pushed the limits on everything and always have won. Sometimes you lose and this is that time. However upsetting it is to lose such a beast of a CPU i know you wont let it beat you down and stop overclocking.

Thanks for sharing, being honest about it all.

Good luck with the RMA. Intel sucks for it but Michell Chan @ Intel China is whom i deal with and she's never let me down on anything.

Keep us posted!

Defo seam 45nm is not a voltage lover at all.