How high of vCore is too high for Core2Duo? 1.35v, 1.40v, 1.45v?

[GundamF91]

My E4500 M0 suppose to have stock vCore of 1.325v, the retail box shows 1.35v Max. But, there's intel's own guide on this E4500 revision M0, which shows that 1.50v is max. I'm wondering what to trust. If ceiling is truly 1.50v then there's more headroom.
http://processorfinder.intel.c...tails.aspx?sSpec=SLA95

Right now I have it running at 1.42v which becomes 1.38 after vDroop, for 333x10 @ 3.33Ghz. I'm running Prime95 stable, full load temp is around 40c now. I am using water cooling (Gigabyte 3D Mercury) by the way.

I'm wondering what's the vCore I should be concerned with, the original number, or the vDroop number?

 

[myocardia]

You should be concerned with what you set in the BIOS. And where did Intel say it's okay to give a C2D 1.5v?

edit: Nevermind, I clicked on the link. I'm willing to bet that whoever typed that page just forgot the 3, between the 1 and the 5.

 

[GundamF91]

Look at the link I had above, it says clearly the voltage range to be 0.85v to 1.5v.

I went bolder and pushed it on. Right now I'm typing this at 333x11, 3.66Ghz, this is the highest I've ever hit. Vcore at 1.6v though. Idling at 30c, and load is topping 60c.

 

[SerpentRoyal]

VID +10% is a safe number (1.46). Under load, the CPU will probably run at 1.42V. If C1E and EIST are working, then the CPU will always see Vdroop.

 

[VirtualLarry]

Originally posted by: GundamF91
Look at the link I had above, it says clearly the voltage range to be 0.85v to 1.5v.

I went bolder and pushed it on. Right now I'm typing this at 333x11, 3.66Ghz, this is the highest I've ever hit. Vcore at 1.6v though. Idling at 30c, and load is topping 60c.

You're brave.

 

[nonameo]

Originally posted by: VirtualLarry

Originally posted by: GundamF91
Look at the link I had above, it says clearly the voltage range to be 0.85v to 1.5v.

I went bolder and pushed it on. Right now I'm typing this at 333x11, 3.66Ghz, this is the highest I've ever hit. Vcore at 1.6v though. Idling at 30c, and load is topping 60c.

You're brave.

I know. That is really high. Most I hear of is around 1.5.

 

[GundamF91]

Yeah, but I guess that's what it takes to hit 3.66Ghz. I don't think E4xxx were made for this high speed, hence it takes a lot of vCore to get it running above 3.6Ghz. I've cranked it down back to 3.33. Everything's okay.

 

[MetaDFF]

Originally posted by: myocardia
You should be concerned with what you set in the BIOS. And where did Intel say it's okay to give a C2D 1.5v?

edit: Nevermind, I clicked on the link. I'm willing to bet that whoever typed that page just forgot the 3, between the 1 and the 5.

Maybe I'm seeing things too, but the last time I checked the VID range for my E6750 SLA9V it was stated as 0.85 - 1.35 V. Now when I check the Processor Finder link for my processor it says VID of 0.85 - 1.5 V. Did Intel update their documents? Or maybe I'm not getting enough sleep...

 

[aigomorla]

Originally posted by: MetaDFF

Originally posted by: myocardia
You should be concerned with what you set in the BIOS. And where did Intel say it's okay to give a C2D 1.5v?

edit: Nevermind, I clicked on the link. I'm willing to bet that whoever typed that page just forgot the 3, between the 1 and the 5.

Maybe I'm seeing things too, but the last time I checked the VID range for my E6750 SLA9V it was stated as 0.85 - 1.35 V. Now when I check the Processor Finder link for my processor it says VID of 0.85 - 1.5 V. Did Intel update their documents? Or maybe I'm not getting enough sleep...

WTF.... that doesnt make sense... 1.5V is way too much on any processor.

And if were talking about intel stock heat sinks, maybe they assume the cpu will throttle 24/7 but the owner will be happy with the super high overclock numbers?

On Air i think the max safe voltage with a tuniq or TR120 is around 1.4-1.42 depending on ambients.

On water the max safe is debatable somewhere near 1.5-1.55 But i dont like pushing 1.5 on my cpu's.

Even if your keeping the heat down on the chip, the voltage would wear out the chip like crazy...

 

[GundamF91]

If someone at Intel did forget the "3" between 1 and 5, then that's a major slip up on their end. Then again, I can't imagine them messing up on all of the spec. It'd appear that all the latest revisions have the 1.5v max rating. This may have something to do with the increased temp threshhold on the latest revisions.

 

[BonzaiDuck]

I pretty much agree with AigoMorla.

There's a reason for the Intel "retail-box" displaying the "maximum voltage" spec. It is probably a good indication of a high-side in a distribution of completely "safe" voltages. That way, the company doesn't fret so much about costs for servicing their "3-year-warranty." Looking at it the other way, it would be the lowest point in a distribution of voltages with increasing risk of failure.

I've found -- could be my imagination, but I try and watch these things -- that there's a point where more and more voltage "notches" give less and less over-clockability. This also seems to follow a pattern if more and more noticeable increases in load temperature. With water-cooling, you can mitigate the heat problem, but there wouldn't be much of an indicatiion about how much electrical stress is being put on a consumer's single processor. Intel would've designed and tested this factor, so they'd have a clearer picture of it.

Last time Aigo posted his screenies, it looked as though 1.46V was his VCORE setting with a superb water-cooling setup.

I see that the resellers report this "0.85V to 1.5V" range, but my guess is that anything above 1.45 -- moving toward 1.5 -- is pushing it. You're sure to generate more heat, and it increases the likelihood of electron migration over time.

 

[JAG87]

1.35 vcc is the maximum safe vcc according to intel. their definition of maximum safe is "a safe and reliable operations for years to come"

1.55 vcc is the absolute max according to intel. their definition of absolute max is "a safe but non-reliable operation".

anything above that, they state that its unsafe, and unreliable. read the white papers folks.

 

[kb2114]

I've gotten mine to run completely stable at 3.6 GHz with a voltage of 1.575 using a Tuniq Tower. I only leave it at 3.6 when I want to run benchmarks or impress people though, I don't feel comfortable with it there for long periods of time. Normally I run it at 3.2 with a much safer voltage of 1.375.

 

[BonzaiDuck]

I agree with JAG87, in that there would be a distribution of voltages and a distribution of processor life-spans at various voltages.

KB2114 is probably prudent in his practices, at least at those settings.

 

[LOUISSSSS]

got mine at 1.375 BIOS right now cooled with freezer pro and it hit 69C with overnight prime95

 

[GundamF91]

Is yours running 3.4ghz as in sig, or higher?

 

[spinejam]

Well, with all of the overclocking peeps in this forum and at XtremeSys...com., how many of us have fried an intel core2 duo / quad to date? How many people have reported an overvolt-failure? Curious?

 

[LOUISSSSS]

Originally posted by: GundamF91
Is yours running 3.4ghz as in sig, or higher?

mine is clocked at 3.55ghz right now...

 

[tenax]

what i'm curious about is is it the voltage..or the operating temperature that kills? i believe it's the voltage unless the temperature is extremely high. (like years ago when i forgot to put the heatsink on my athlon 800 and in 15 seconds it was literally toasting.

i have been running my 6400 at 8 x 425 with a big typhoon in an htpc case for 1.5 years and aside from reboots, 24/7..not pushing it really hard in anyway (haven't run a game on it..all htpc, web, email, documents). it runs 51C at idle on both cores at 1.25 volts..so actually undervolted..it will get to the 60C range if i push it..never a glitch or a sign of problems. and to an earlier post on this thread, yes, in spite of overclocking like crazy..who has heard of any frying a dual core? not me.

 

[MetaDFF]

Originally posted by: tenax
what i'm curious about is is it the voltage..or the operating temperature that kills? i believe it's the voltage unless the temperature is extremely high. (like years ago when i forgot to put the heatsink on my athlon 800 and in 15 seconds it was literally toasting.

Well both a high voltage or a high operating temperature can kill a chip in different ways.

If the operating temperature is too high, the current density through the metal interconnects can cause the metal to be "burned off" in a process known as electromigration. Most chips today are designed to handle the current density up to 125 C before electromigration starts to occur. Of course if your heatsink isn't on the chip when you power it on, the rapid heating of the die can also cause it to crack.

Even if you cryogenically cool your chip and temperature does not become a factor, an extremely high voltage can still cause chip failure. An extremely high voltage can cause the gate oxide in the transistor to experience dielectric breakdown. Since the gate oxide thickness in today's chips are literally 10 atoms thick, it is very easy to destroy this gate oxide layer. Once this oxide breaks down, your transistor is "blown" an can no longer act as a switch, so your chip also fails.

 

[tenax]

great information meta..and on that basis, as i suspected, i would be more concerned about excessive voltage to my chip, than i would be heat. there is no way my processor will get anywhere near 125 or 90C for that matter..hot? yes but i'll take my chances on a 60C processor on 1.25 volts and not sweat it rather than 1.5 volts applied constantly threat to the cpu

 

[Tempered81]

i keep my vcore at 1.38 to 1.39 and really cant boot into windows higher than 3.35ghz with that voltage. my temp is too high at this point 55 idle on stock air to worry about ocing higher, then really no need for more voltage at that point. i hear its wise to stay below 1.4volts. hope that helps.

 

[Pabster]

1.6V? Man, that's living on the edge. :laugh:

 

[myocardia]

Originally posted by: MetaDFF
Well both a high voltage or a high operating temperature can kill a chip in different ways.

If the operating temperature is too high, the current density through the metal interconnects can cause the metal to be "burned off" in a process known as electromigration. Most chips today are designed to handle the current density up to 125 C before electromigration starts to occur. Of course if your heatsink isn't on the chip when you power it on, the rapid heating of the die can also cause it to crack.

Even if you cryogenically cool your chip and temperature does not become a factor, an extremely high voltage can still cause chip failure. An extremely high voltage can cause the gate oxide in the transistor to experience dielectric breakdown. Since the gate oxide thickness in today's chips are literally 10 atoms thick, it is very easy to destroy this gate oxide layer. Once this oxide breaks down, your transistor is "blown" an can no longer act as a switch, so your chip also fails.

Actually, if you want to know what electromigration is all about, read this thread: http://forums.anandtech.com/me...ey=y&keyword1=CTho9305 The second post in that thread was written by someone who works for either AMD or Intel.

 

[BonzaiDuck]

See, I don't think INtel (or AMD, for that matter) need "us" to test their processors.

The "need us" to buy them.

But they have the lab resources to determine the statistical probabilities of failure at certain levels and kinds of stress.

My view of it -- we're sort of "outlaws" -- playing outside the parameters the manufacturer requires to honor its 3-year-warranty. From that view, there aren't any "hard and fast rules" for us to follow, other than to willingly "own" the responsibility for our hardware failure if it indeed occurs.

So I think you can artfully over-clock your system and feel accomplished at it by staying within some prudent limitation.

By analogy, let's suppose you want to get from Bakersfield to Sacramento as soon as possible on Interstate 5. You know that 20 mph over the speed limit is reckless driving, and you might be stopped, issued a ticket with a fine, and even asked to appear in court. Everybody else on I-5 is "speeding," but the CHP doesn't stop anyone going less than +10mph over the limit -- and they just sail on by, and eventually arrive in Sacramento without incident, in some optimal time-frame.

The other approach is to assume you posi-tutely, abso-tively hafta get to Sacramento from Bakersfield in 2.5 hours. That's like "I HAFTA over-clock this 2.4 Ghz processor to 3.8 -- regard-lesss!!!" If you're driving, the odds of getting a ticket -- even getting thrown in jail with your 2.5 hour requirement, are pushing 100%. You're also going to burn up your engine, and maybe kill a few people.

At least, Over-clocking is not "life-threatening."