D@mn I'm enjoying this QX9650.

[NuT]

How's this looking guys?


http://img219.imageshack.us/img219/7966/3808au7.th.jpg (http://img219.imageshack.us/my.php?image=3808au7.jpg)

[[XC] NetburstXE]

Can you post a screenshot with Coretemp or some other temperature monitoring tool open? Anyways, good results, but could probably go higher.

[fornowagain]

You need as much as vcore as that? Small ffts will get it hotter btw.

[CCC]

You need as much as vcore as that? Small ffts will get it hotter btw.

pretty normal from what ive seen.. personally ive had 2, the first one needed 1.42v for 4200mhz.. the one i have now needs 1.42v for 4000mhz

[road-runner]

The one I just got needs to be set on 1.425 in the bios for 3.9Ghz...

[NuT]

Sorry for the sh1tty image guys.

According to core temp my tems were 56,54,53, and 52.

And with UncleWebbs real temp they are exactly 10 degrees lower!

Would it be safe to run theses settings 24x7?

Thanks!


It got up to 61c while running prime, not gaming.

[[XC] NetburstXE]

It would be a good idea to lower the voltages if you are going to run it 24/7. In excess of 1.4 volts will eventually kill a 45nm chip. However, if you are only planning on keeping it for 2-3 years, it will be fine. It will take longer than that to totally kill it.

[WFO]

It would be a good idea to lower the voltages if you are going to run it 24/7. In excess of 1.4 volts will eventually kill a 45nm chip. However, if you are only planning on keeping it for 2-3 years, it will be fine. It will take longer than that to totally kill it.

Just out of curiosity, do you have a link to info supporting the 1.4 will kill a 45nm chip? I read in another thread it took in excess of 1.450. :shrug:

[CCC]

Just out of curiosity, do you have a link to info supporting the 1.4 will kill a 45nm chip? I read in another thread it took in excess of 1.450. :shrug:

i ran my first one at 1.42v for 24/7 and it died.. having said that though while running benchies i had it as high as 1.60v just for a few runs but temps never got higher than 65C.. Not sure why it died.. might not have had anything to do with voltages.. I just cant imagine 1.45v or around there would actually kill these chips(assuming good cooling) but i guess every chip is different

[joe.denice]

thats a decent OC, I had my QX9650 at 4ghz with 1.33 or so volts but for some reason they chip got really really hot and I noticed even at stock this was an issue, come to find out it was the chip and not my setup, but already swapped back to the maximus formula. Right now though I am playing with an e8500 w/ 1.275 volts doing 8x500 for 4ghz.

[antari]

Are you sure you need that much voltage?

I admit this screenshot is only for 1hr quad-prime, but the machine has run F@H for many days at this setting stable

http://img518.imageshack.us/img518/5581/p954ghzjm1.png

[NuT]

Are you sure you need that much voltage?

I admit this screenshot is only for 1hr quad-prime, but the machine has run F@H for many days at this setting stable

http://img518.imageshack.us/img518/5581/p954ghzjm1.png

Yeap it needs that much, u just have a magic chip. :up:

[antari]

Yeap it needs that much, u just have a magic chip. :up:

Thanks :up: I guess even though yours need more volts it seems to run a lot cooler than mine. I think I need to upgrade my cooling.

[CyberDruid]

I would go easy on the volts. I killed one with too much PLL.

Anything over 1.4 is asking for trouble. Remember the Northwood Sudden Death Syndrome about 5 years ago? Same thing.

Everything will be looking ducky and then poof.

I keep mine below 1.4...do some research you'll see it's pretty common.

[Lestat]

I would go easy on the volts. I killed one with too much PLL.

Anything over 1.4 is asking for trouble. Remember the Northwood Sudden Death Syndrome about 5 years ago? Same thing.

Everything will be looking ducky and then poof.

I keep mine below 1.4...do some research you'll see it's pretty common.

PLL voltage is sucha huge variable.

some boards take more, some cpu's take more. its never the same.

cpu pll is not cpu vcore. neither is it cpu vtt they are all 3 different voltages.

pll is your overclocking chip, and thus the voltages given to it to allow higher fsb (as i understand it).

cpu vtt, well, im still at a loss at what it's true purpose is. and vcore is just that.

there is also the clockgen voltage, which could be misconstrued with pll voltage. but pll and sb voltages are linked.

personally i would be more worried about CPU VTT than pll. since cpu vtt is directly linked to the cpu and cpu vcore.

[JumpingJack]

pll is your overclocking chip, and thus the voltages given to it to allow higher fsb (as i understand it).




Actually, the PLL is not a seperate chip. PLLs are built into the chip and are used to set the clock for the CPU. You may be thinking of the clock generator on the motherboard (the major one), from which most all other clocks are derived either by dividing the clock or setting via the phase lock loop.

http://www.realworldtech.com/page.cfm?ArticleID=RWT051607033728&p=8
Here is a nice picture showing Barcelona's PLLs.

http://download.intel.com/design/processor/datashts/31872602.pdf
Pin D23 and C23 bring the PLL voltage and clock signal into the CPU package.
(Figure 4-1)

Jack

[Leeghoofd]

Yeap it needs that much, u just have a magic chip. :up:

Indeed your Chip is average, I have one too that needs 1.39 volts now for 4ghz stability (it used to be 1.37 a month ago so it's degrading for sure) but this is my bench chip and had 1.6 volts running through it for short bench periods... other one I got does 4ghz at 1.26volts and that one I run it daily at 4ghz but not one Mhz higher as I'm a bit scared of loosing it...

Also scared for the mad PLL voltages of my Maximus Formula as it constantly overvolts it... one of the reasons I sold it...

[fornowagain]

pretty normal from what ive seen.. personally ive had 2, the first one needed 1.42v for 4200mhz.. the one i have now needs 1.42v for 4000mhz

Just sounds a tad on the high side to me, I thought mine was bad when it needed 1.38v for 8hrs prime small ffts at 4Ghz.

[GAR]

nice if its stable for 24 hour use.

[fornowagain]

how is that high 1.35v is less than the max default voltage? which is 1.36v

Actually its not. You are referring to the voltage identification or VID, that's not the vcore. If you actually read the specification you will see that maximum process voltage is 1.45v, intended for transient conditions.

Continuous maximum values are somewhat lower.

http://img366.imageshack.us/img366/876/dscf2392uf8.jpg (http://imageshack.us)

In fact what I was saying is 1.4v+ seemed a lot for only 4Ghz, when I've seen a lot do better and I considered 1.35v as mine decidedly average.

[Leeghoofd]

How is a chip degrading when your running it only .03v over the max default voltage ? No way maybe you used a bios update that change the voltage you needed

First up my Vid is 1.125, CPU doesn't run even close to that on stock speeds (9 x 333 ) on 3 different boards all with latest bios...( Asus P4K Dlx, EVGA 780i and Foxconn MARS) This is my bench CPU so it has endured (under a high quality WC kit and pretty low ambient temps) some beefy voltages (1.55-1.6) and run at and above 4.5Ghz...for short durations...

At start ( before the benches and stuff) the CPU required 1.16volts to be stable in prime (bummer not good)... now it needs almost 1.19volts for stock speeds... same for 3.6ghz ( which I run this one on 24/7) used to be 1.25V now 1.29V, 4ghz used to be 1.37 now 1.39 volts... CPU is slowly going downhill...as I will continue on this one with benching till it pops.
I'm reffering to degrading as the CPU had done the same speeds before with lower voltages : rock stable, now it even BSODS on me if I don't give it enough juice...First like you claimed I suspected the mobo to be the issue but CPU did the same thing in another mobo. This made me very suspicious...

My other Qx still does 4ghz at 1.26volts , no worries with that one... but it has never got any voltages higher than 1.4...

And to be honest if you buy an extreme CPU, who's gonna run it at stock speeds and voltages mentioned on a stupid box ? (I think it said 1.25 v on my box)...or data sheet...? These CPU's should be more durable than this... surely for that money... and on a note Intel replied to me I should use the retail heatsink for any use or abuse of their CPU's... yeah right :rofl:

[fornowagain]

those are not the vids those are the max default voltages. I know this because in the other specs like on a Q6600 the max voltage is 1.5v and thats not the vid.

Ninja.

Those are the vids, voltage ranges, as in min/max voltages ranges for your chip, default vid or stock vid will be something like 1.25v, but what thats is, is intels max recommend voltage or vid you can manually set, those are the Min/ max default voltages ranges ( in the case of the example a Q6600 .85v to 1.5v. I know this because in the other specs like on a Q6600 the max voltage is 1.5v and thats not the stock vid, no chip is auto set to that by intel.


Wrong on both counts. Here's an idea, go and read it. You know what VID is right?

http://download.intel.com/design/processor/datashts/31872602.pdf

http://img329.imageshack.us/img329/5521/qxyt5.jpg (http://imageshack.us)

http://img359.imageshack.us/img359/904/vcchw6.jpg (http://imageshack.us)

Oh and btw, the maximum process value for transient voltage on a Q6600 is 1.55v.

[fornowagain]

Also the max voltage on the Q6600 on the box I own is 1.95v and you think thats still correct? it not the default voltage set by the chip is higher than that. A Qx9650 can take 1.45v on good air cooling without problems, don't just make a statement show me some evidence it can't.
Is that right, must be a one off, my Q6600 box said max was 1.35v. I'm not making speculative statements on what a QX can 'take'. I'm correcting you and providing what Intel actually SAY is the maximum. Nothing more.

http://img356.imageshack.us/img356/6222/dscf1769bs6cr3.th.jpg (http://img356.imageshack.us/my.php?image=dscf1769bs6cr3.jpg)

Max VCC is always lower than the VID from Voffset and Vdroop.

http://img80.imageshack.us/img80/5269/intelcore2quadprocessorax8.png (http://imageshack.us)


okay and whats that mean? i can only go with what intel makesd available and my posting is just as correct as yours. Look here
As correct? err no. Do us all a favour and go read the subject...carefully.

Once again, VID not vcore. Like I said, you know what VID is right? Anyways, its a load line (depends on EIST etc) factory set within a range for the individual cpu thermal profile. The 'value' you see in coretemp etc ain't the stock vcore, not even close. In short, VID defines the maximum TRANSIENT voltage (peak overshoot or spike if you like) for VCC, it is not the default or maximum Vcore. Further than that is the absolute maximum process voltage. That's the highest transient voltage they say the cpu can 'survive', but with a degraded lifetime (see 2.6.1 most docs).

[NuT]

I'm gonna run 1.43 vcore on this puppy, will let you guys know what the out come is. lol


I'm not scared!!!!!

[fornowagain]

You shouldn't be unless you have a defective chip. Voltage almost never kills a chip heat does. As long as your load temps stay under 60c your fine. High Voltage will degrade your chip over long periods ,however, I have owned the follow chips over the last two year alone,3600x2 , 3800x2 ,6000+, E4500,e6400,e6420,e6600, e6700,e6850,Q6600, X3210, X3220, E8400 and now QX9650 and I have pumped as much as 1.75v during benching thru these chips and never had one die. I have hit temps on the tjunctions as high as 83c and none have died. For prolonged use most people say stay under 1.5v on the core and depending on who you talk to under 50c or 60c for load temps if possible. I'm not new at this and have belonged to another forum, posting thousands of times. If this Qx9650 voltage death thing was a issue , I would have heard of it and read about it , and I haven't. People would be going crazy if 1,000+ chips were dieing as you say they are. I have seen a guy who seems to spout(post) voltage data but has yet to show me one link on a tech page,one news article or fact sheet back up what he says is a known problem for QX9650 chips so once again backing up his statement with facts. Show me this epidemic of voltage killed Qx9650's. If your right I will be the first to say i'm wrong but as far as I know I'm not.
You'd think with all that experience you'd actually know what you're talking about. You seem to have a profound lack of knowledge. And you spouted the voltage data first, lol.

Where did I say any were dying? What known problem? Lmao, I haven't said a word on any of that! What are you talking about? I have merely corrected your misinterpretation of Intels design specification with regards to maximum core voltage. I have not commented on what I think is a safe voltage. Everything I have said on specification is easily verifiable statements of fact. I've even given you the link to docs with screens, as you don't seem inclined to read them yourself. Jez some people:shrug:

[jonny_ftm]

Where is the statement more than 1.40v will kill 45nm chips coming from. Thats only .04v higher than the max default voltage of 1.36v? What source is that from most intel chips can take atleast 1.50v. i'm not saying your wrong but where is this from?

Hey, take your sward away, get some knowlodge before you come back instead of throwing :banana: and confusing noobs

Q6600 G0, as of intel recommandations should not be given more than 1.5v
QX9650 shouldnt be given more than 1.3625v

You and most forget about overshooting voltage implemented by vdroop intel specifications. For us, motherboard manufactures decided to get rid of it, if you wanna take risks with your chip. If you set bios to 1.3625v while removing you actually overshoot your chip by up to 0.09v on quiet all motherboards. This is a spec design you can't get rid of. So, even if your excited by your multimeter and CPU-z showing you run 1.36v, you're actually shooting it by up to 1.46v. Those spikes, the chip gets them thouthands times a second. Thus, if you were open to some knowledge instead of playing it like a silly stubborn, you'll understand how much skilled people than you got their chips degraded as low as 1.35v

Now, as we got your useless explanations, no need you come back with same things. Better get some real knowledge or at least be more modest in your pretentions

[jonny_ftm]

You shouldn't be unless you have a defective chip. Voltage almost never kills a chip heat does. As long as your load temps stay under 60c your fine.

:rofl:

Hey for all, keep in mind that forums are full with such sh*it and be carefull with your CPU

[Leeghoofd]

Trt 740 I think you can not compare 65nm with 45nm chips... and indeed they are not dead ( yet ) but there are numerous reports of deaths by the big clockers that use extreme cooling...so heat musn't have been an issue at -80°C. My CPU's aren't allowod to go over 65°C, if they do I switch off...

But some will degrade with voltage as they seem not as resistant as the 65nm counterparts... there are lesser ones out there and I am an owner of one :)

You can discuss with all the theories in the world...and many like to see stuff black and white on paper, remember that there are always exceptions to the rules!!

I see what I see here and have been long enough in this business to notice some weak spots... some survive 1.4+ volts for 24/7 use but I'm pretty sure some will degrade rather rapidly ( I did not say DIE )

Trt70 I just hope you get a good one and can enjoy it long time... as they cost heaps and sadly this does not reflect in realiability and build quality... Shame on you Intel !

[Richard Dower]

OK...y'all scaring me with vids and voltage.

[fornowagain]

You right chip death and degredation was never mentioned on this thread, I made it up.
I certainly didn't mention it and yes you made up that whole 1000 chip dying rant.


I was under the impression this thread was about safe voltage levels for a Qx9650 and that sudden chip death was being indicated on this thread. It's now clear to me that you didn't read the thread at all before you posted.
Haha, try again. I posted third and commented on the level of voltage needed for stability, I haven't deviated from that. In fact it was you who mis-read me to start with????


You were to busy explaining a bunch of useless nonsense about chip voltage , that would have confused more noob then I ever could. All I said is generally 1.45v is considered safe on most Intel chips. a rule of thumb and load temps below 60c are also considered safe buy most overclocker. I also then listed intels max voltage recommendation from their site.
I was talking to you, you're no noob allegedly. I merely corrected noob errors and what 'you' consider safe, especial in light of your obvious lack of knowledge is ludicrous. And you posted Intels VID's for crying out loud, read something for god sake.

I hardly think any information I posted was wrong or confusing. What I posted is hardly a new. Degradation usually occurs from extreme volatge and high heat over long periods of time.
I'd say that's the problem, you hardly think. All of it was wrong and misleading. Unfortunately these chips ARE new, no one really knows the long term effect of high voltage on the high-k dielectric. Degradation has quite a few mechanisms, 'usually' doesn't apply as yet. Take a moment to search the Wolfdale thread for degradation.


However, nothing I posted here will get you to that point. I will say nothing you have posted would help anyone trying to overclock a Qx9650. It might scare them or confuse them.
Hopefully most noobs would have the good sense to actually read the Intel documentation and take the time to understand it, rather than post pointless mis-quotes.

I think you should explain all the useless Vid knowledge to them it might help them understand and explain the answer to this thread topic. You are a true font of knowledge and a true tech giant.
Ah, I see we have found your level. So be it.

The document again, READ IT > http://download.intel.com/design/processor/datashts/31872602.pdf

Maximum casetemps are 64.5C (60C where did you get that from?), obviously coretemps will be higher. And this is before assertion of the throttling point TCC, not a particularly dangerous place for the cpu. My QX9650 has a stock VID of 1.0375v, it actually helps if you understand the principle of load lines and VCC. However look at figure 2.1 on the docs, there is an offset voltage for Icc (load states). This is 'how much' lower the VCC (core) should be for static values. Static values are the ones set after Vdrop (or Voffset) & Vdroop. The MAXIMUM VID is factory set, so my QX would have maximum vcore at stock of 1.0095v (peak 1.0595v < 25μs) or there abouts.

Now get this, its not 1.36v as you keep saying, that's not even VCC. That's the maximum VID for the generic cpu range, not THIS chip. And its not the 1.45v, that's the maximum voltage for transient spikes that Intel say WILL cause damage.

Now if you want to go higher than the stock voltage, fine never said not too. Just understand what the stock value actually is, be aware. In my example your safe 1.45v is a 43% overvolt from stock. Good, bad? That's up to you.

http://img187.imageshack.us/img187/4686/transientvdroopza1.jpg (http://imageshack.us)

These a good article on the subject here......it has pretty pictures.

http://www.anandtech.com/showdoc.aspx?i=3184&p=3


The CPU VID setting establishes the absolute maximum allowable processor supply voltage experienced during transient conditions and is not the target idle voltage. We hope this statement draws attention to this important distinction, as many believe the opposite to be true - a mistake all too commonly made. Together, Vdroop and Voffset ensure that the peak CPU supply voltage seen during heavy to light loading changes remains well below the established maximum. If you determine that 1.17V, as in the case above, is not sufficient for maintaining CPU stability under load, simply increasing the CPU VID does correct the problem. Let's now examine how the system responds if we remove Voffset.

Personally I'm happy to put a few volts through it (water), still primes at 1.38v at 4GHz (so far fingers crossed).

http://i180.photobucket.com/albums/x32/fornowagain/th_cpu_z_48.jpg (http://i180.photobucket.com/albums/x32/fornowagain/cpu_z_48.jpg)http://i180.photobucket.com/albums/x32/fornowagain/th_QX40.jpg (http://i180.photobucket.com/albums/x32/fornowagain/QX40.jpg)

[Leeghoofd]

Trt stop posting all this extra posts, do a multiquote then plz.. ofcourse there's no 100&#37; proof...and everyone is entitled to it's own opinion...
But if a CPU did 4ghz daily at 1.36 volts and needs now almost 1.4volts to be stable again this surely must ring a bell, doesn't it... and voltage max I gave it was 1.6 ( in bios no loadline calibration and co enabled, so it was around 1.56 real ) and it is hooked up to a 120.3 thermochill, D-Tek Fuzion and a LAING ultra pump... All benchmarks were done at nite during cold belgian winter nites with the window open (I know I'm a geek)... so pretty low ambients around 5&#176;C
Voltage might have been a bit extreme but temps were A-okay during all sessions... yet it degraded... and I'm not the only one on that matter, we are with 4 already that have this same issue... and those guys don't pump 1.6 volts or more into their CPU's as they don't bench at all... they just game and have really beefy cooling equipment like many amongst us here...and they aren't first time builders either....

You can tell me anything you want with magic numers, tech sheets,etc... there are degraded CPU's out there and why did that happen ? Heat, Voltage ? You tell me... but it is comon knowledge that these CPU's react totally different to voltage and heat then what we were used too... many will complain in the coming months... mark my words... and 1.4 volts daily might be just in the safe zone but I wouldn't recommend it to anyone...

Let's get back on topic and be happy with the thread starters CPU :)

[antari]

trt740, you seem awfully happy to ensure people that voltages in excess of 1.40V, or even 1.36V are safe. Are you going to replace the CPU of someone who runs 1.42V daily after reading your posts in this thread, and has their CPU degrade badly? Easy to make a stand when it's not your chip that will get hurt, ey? :rolleyes:

[trt740]

trt740, you seem awfully happy to ensure people that voltages in excess of 1.40V, or even 1.36V are safe. Are you going to replace the CPU of someone who runs 1.42V daily after reading your posts in this thread, and has their CPU degrade badly? Easy to make a stand when it's not your chip that will get hurt, ey? :rolleyes:

No sure won't, i'm done good luck appears this topic is not open for discussion, don't use more voltage than recommended by intel or your chip might die. Don't overclock your chip it voids the warranty and there is a bunch of risk involved. Just run it at stock and all will be well. Some serious mean people on this forum WOW!! It has been lots of fun being talked down to. :up: To the voltage , cpu designer fornowagain thanks again for not being arrogant and flat out condicending. I'm not sure how anything you have posted help this thread. Apparently my impute was not needed or wanted so I have cleaned this thread. good luck with your chips.

[lior307]

I didn't see in my whole life a Dead CPU of intel .
I have also this CPU and i put 1.84v to this and nothing happend
im using it with 1.74v for about a month .
Everything works just fine .

[Leeghoofd]

With Subzero cooling you have ofcourse way more headroom and require less voltage for the same speeds...sadly not reflecting the cooling of the majority off the masses... as many pump voltages through these CPU's with average cooling...

You use that Phase 24/7 sir or just when benching ? Kindda interested in getting one myself, though pretty hard to find in Belgium...