Typical Ivy Bridge Overclocking Results

[Spaceme1117]

Hello. I just recently purchased an Intel Ivy Bridge i7 3770K. I have it water cooled using Indigo Extreme TIM. My temps at idle are 35-45 at 4400 mhz and 1.280 vcore. How high should I be able to go? Also, what would be the highest safe vcore that I could push? Thanks.

 

[Ken g6]

Idle temps don't matter. What are your load temps? I think most people are limited in their voltage by load temps more than by anything else - except for those who de-lid or adjust the lid.

 

[smakme7757]

Load temps with Intel Burn Test are best. Prime 95 does't generate nearly enough heat or stress on the CPU.

Anyhow 4.4 to 4.5 is a nice overclock on Ivy.

 

[aaksheytalwar]

1.25-1.3v is the max I would like to push any ivy for long term and stable usage. Keep load temps below 80c or so.

 

[hokies83]

1.52 is MAX Intel Spec.. i run 1.55v I do not plan to keep the cpu for 5 years... and want the most out of my Chip.

 

[aaksheytalwar]

It won't even last 6 months at 1.52V

 

[sm625]

1.52 is within intel's spec, why wouldnt it last 6 months? I bet it would run for years at 1.55. (Maybe not at the same max clock, but I bet the chip still works 5 years from now.)

 

[Idontcare]

1.52 is within intel's spec, why wouldnt it last 6 months? I bet it would run for years at 1.55. (Maybe not at the same max clock, but I bet the chip still works 5 years from now.)

1.52V for a 22nm IB is within Intel's spec!?

1) I completely doubt that, but
2) link to said spec?

 

[Kenmitch]

Wasn't this chart floating around when IB was released. I don't think it's official Intel chart but does ref Intels max voltages.

IB's overclocks are dictated by temps more than voltage as IDC has shown. Even bare core the temps get crazy high under heavy loads.

 

[l2ez4m]

So they say, but honestly it doesn't look very realistic

 

[jvroig]

Last I remember, there was a thread here where 1 or 2 actual Intel engineers (the known Intel employees we have around) actually mentioned that 1.35V is not safe, although they wouldn't say what was safe exactly for obvious reasons (accountability to their employers).

So 1.52V? No way.

I'll see if I can find that thread using my google-fu.

 

[jvroig]

It was this thread: http://forums.anandtech.com/showthread.php?t=2243736

"MrDudeMan", posts #24 and #32, then "PM" posts #80 and #83 (sorry for the non-links, I'm lazy).

Highlights:
-post #32 states 1.4V is already "nothing short of retarded" for Sandy.
-post #80 states "I look at some of the voltages that people say are "safe" and I think that I'd be worried about doing that with a chip that I am playing with"
-post #83 clarifies why they won't commit to a safe value.


So again, I repeat: 1.52V? No way. Not for Sandy, and definitely not for Ivy.

 

[Kenmitch]

It was this thread: http://forums.anandtech.com/showthread.php?t=2243736

"MrDudeMan", posts #24 and #32, then "PM" posts #80 and #83 (sorry for the non-links, I'm lazy).

Highlights:
-post #32 states 1.4V is already "nothing short of retarded" for Sandy.
-post #80 states "I look at some of the voltages that people say are "safe" and I think that I'd be worried about doing that with a chip that I am playing with"
-post #83 clarifies why they won't commit to a safe value.


So again, I repeat: 1.52V? No way. Not for Sandy, and definitely not for Ivy.

IB overclocking is dictated by temps more than voltages. It's not like you can run 1.5v's in one without some heavy duty cooling anyways. Well I guess you could if you just want your chip to throttle down anyways but then what would be the point.

For the faint of heart there is always a solution. In this case it's the Intel protection plan which is pretty cheap for what it offers.

http://click.intel.com/tuningplan/

 

[jvroig]

Yes, yes, that's all correct but beside the point.

The contention was 1.52V was official max Intel spec. I distinctly remember that that was not the case, hence I cited that thread, where 2 actual engineers from Intel chimed in. End of contention.

So regardless of whether you plan to cool it or not, or have a reasonable plan to do so, the earlier claim (even before you joined the thread and posted that image table you have) that 1.52V is max Intel spec is completely not true.

 

[Idontcare]

Wasn't this chart floating around when IB was released. I don't think it's official Intel chart but does ref Intels max voltages.

IB's overclocks are dictated by temps more than voltage as IDC has shown. Even bare core the temps get crazy high under heavy loads.

That is nothing official. I can label a chart as "official Intel spec" too, doesn't make it so.

So they say, but honestly it doesn't look very realistic

That is VID, voltage identification, a pre-programmed lookup table. The range on the table has nothing to do with the safe voltage operating range itself.

VID is not to be confused with Vcc. Vcc is the CPU operating voltage.

For the faint of heart there is always a solution. In this case it's the Intel protection plan which is pretty cheap for what it offers.

http://click.intel.com/tuningplan/

Unfortunately even the tuningplan warranty is null and void if you lap or delid your CPU. Any physical alteration of the CPU or the CPU package nullifies the warranty.

So for folks who want to delid or lap, but still would like to see their CPU last >6 months, they will still want to know what is a practical upper limit to the operating voltage because they don't get a second chance to get it right without forking out another $300 for a replacement CPU.

 

[hokies83]

1.52 came from Intel it is in the De Lid Club from OCN

Ive Been running 1.55v for about 70 days with out issue.. 30hrs of stress testing 100hrs of gaming 200 hrs of video encoding...

And please do not throw SB Voltages and compare that to Ivy...

Ivy was heat limited SB is volt limited..

A de lided Ivy is ready to rock n roll.

I mean example of how much more Ivy can take voltages ...

Try to do this Voltage on a SB cpu on a 99$ mb and see if it is still alive afterwards
This was done on an Air cooler..

 

[Idontcare]

1.52 came from Intel it is in the De Lid Club from OCN

If that were true, that Intel officially published their spec value for max IB voltage, don't you find it the least bit odd or suspicious that the only place they elected to publish it at was a minor subforum on a 3rd party website?

Do you think they just forgot to publish it on their own servers within their own technical specs? Or that they forgot to send out the same memo to all the other hardware review sites?

C'mon man, use a little common sense here.

 

[hokies83]

If that were true, that Intel officially published their spec value for max IB voltage, don't you find it the least bit odd or suspicious that the only place they elected to publish it at was a minor subforum on a 3rd party website?

Do you think they just forgot to publish it on their own servers within their own technical specs? Or that they forgot to send out the same memo to all the other hardware review sites?

C'mon man, use a little common sense here.

Our Guy in the De lid club Contacted Intel Directly..

The Intel Tech brought up there engineering charts..

And Said 1.52 Max Intel Spec.

sheet 1, page 84, 7.10.1, Table 7-4
http://www.intel.com/content/www/us/en/processors/core/3rd-gen-core-desktop-vol-1-datasheet.html

 

[Ferzerp]

That's just the range that the register that reports Vid is capable of presenting (Exactly like IDC suggested earlier)....

That has *nothing* to do with "max Vcc"

Think of it like a gauge on your car. My car's speedometer goes to 220, but it can only get to 200 (and only then in the most perfect of circumstances). Just because it can display it doesn't mean it is a meaningful number.

 

[JAG87]

Lol 1.52V

That is not the max Intel spec, that is just the highest VID value that is assignable in the VID table, doesn't mean that it's the highest you can go without damage. Anything over 1.2V is pretty unsafe on Ivy Bridge for long term reliability.

 

[hokies83]

That's just the range that the register that reports Vid is capable of presenting (Exactly like IDC suggested earlier)....

That has *nothing* to do with "max Vcc"

Think of it like a gauge on your car. My car's speedometer goes to 220, but it can only get to 200 (and only then in the most perfect of circumstances). Just because it can display it doesn't mean it is a meaningful number.

Well this is the Above posted Chart by Sin0822 who has a Degree in engineering and is now back at Georgia Tech Getting his Masters in Engineering.

It is not from some 3rd Party web site or some joe blow.. It is from the Most respected " Engineering guy on OCN " and on the Ivy Bridge overclocking Guide Thread in the OP.

 

[Ferzerp]

Umm, that's a 3rd party Joe Blow (and a student no less)..... :/

 

[Kenmitch]

1.52 came from Intel it is in the De Lid Club from OCN

Ive Been running 1.55v for about 70 days with out issue.. 30hrs of stress testing 100hrs of gaming 200 hrs of video encoding...

And please do not throw SB Voltages and compare that to Ivy...

Ivy was heat limited SB is volt limited..

A de lided Ivy is ready to rock n roll.

I mean example of how much more Ivy can take voltages ...

Try to do this Voltage on a SB cpu on a 99$ mb and see if it is still alive afterwards
This was done on an Air cooler..

My chip craps out just above 5.6ghz

Yep....It's still alive if your wondering about it.

 

[hokies83]

My chip craps out just above 5.6ghz

Yep....It's still alive if your wondering about it.

It was the 1.86v i was pointing at :biggrin:

 

[Kenmitch]

It was the 1.86v i was pointing at :biggrin:

My chip didn't seem to respond to vcore after 56x so I just gave up on trying higher.

On another note it takes a lot more to kill a cpu than one thinks. I remember IDC stating he pumped 2v's plus in his 2600k(?) during an Oops moment while playing around with offset voltage and it lives to this day