Sign in with
Sign up | Sign in
Your question

vcore and vdroop questions

Last response: in Overclocking
Share
February 6, 2008 11:25:48 AM

I am running my e4600 @ 3.3 (275x12) with EIST enabled. Because I am not gaming, most of the time I will be at 275x6=1.65. My vcore is 1.41875(Bios), 1.376(CPU-Z). I would like to leave these settings to take advantage of the 1.65 (vs 1.2 stock) even with EIST enabled. My heat is managed OK (load cores at 60-62).

My question is whether simply having so much more vcore (vs stock) is by itself detrimental to the cpu (again, assuming heat is in safe range.)

Re vdroop, what is the significance? Aren't the important parameters how much v the cpu is getting, as measured by CPU-Z, and the resulting heat? Why does it matter that you set your vcore higher in your Bios?

Thanks very much for any advice.

More about : vcore vdroop questions

February 6, 2008 12:03:20 PM

Intel max spec for voltage is 1.50 vCore. As long as you aren't over that you aren't exceeding Intel specifications.

vDroop is Idle->Load. Long story short it prevents the voltage from exceeding the VID. You set the VID in the BIOS, then with vDrop you get Windows Idle (as read with CPU-Z).

What do you mean why does it matter?
February 6, 2008 3:09:38 PM

Thank you, cnumartyr. Most of the messages I have read seem to imply that vdroop is a bad thing (and there is that thread about possibly eliminating it on an Asus board). You seem to say that it is just a fact, neither good nor bad, which is where I end up. This would imply that the 1.5 max spec vcore is as read by CPU-Z. True?
Related resources
February 6, 2008 3:14:13 PM

Yes.

Edit: The reason people want to eliminate it is to run lower voltages.

There are two things.

vDrop (vOffset) which is BIOS (BIOS is also the VID setting) to Windows Idle
vDroop which is Windows Idle to Windows Load

Excessive vDrop can make it harder to achieve a high overclock (if the max Voltage in the BIOS is 1.7v and you have a .1v vOffset than the max voltage you can achieve is 1.6v Real). Some people want to eliminate it for that reason.

vDroop is there to prevent a voltage spike above VID. If you have a power circuit under full load and switch to completely idle it will spike the voltage for a second. vDroop prevents this spike from exceeding VID. It's not a bad thing or a good thing, it's just there.

Removing vDroop can also potentially put too much of a strain on the power circuit and cause burned out MOSFETs if the board was designed with vDroop in mind (which most are).

The reason people remove vDroop is that if you need 1.4v at full load to be stable would it be better (for temperatures) to have a 1.408v Idle or 1.5v Idle? The latter has more vDroop.
February 6, 2008 5:16:49 PM

Again, thank you for this information. Being a first time oc'er, I hadn't even picked up on the difference between vdrop and vdroop, although I had seen both in my oc'ing (I thought the different spellings were just peoiple's different ways of referring to the same thing.) Interestingly, and I think in keeping with what you are explaining, I have noticed that as I have crept up the oc settings for my chip, the settings where there is no vdroop provide a more stable oc. I take it that it is normal that although all settings for me have had vdrop, only some settings have had vdroop.

So with all I have learned, I think I will try for 3.4, keeping vcore in CPU-Z under 1.5, and hope my temps stay OK. At least until the summer.
February 6, 2008 6:26:19 PM

cnumartyr said:
Intel max spec for voltage is 1.50 vCore. As long as you aren't over that you aren't exceeding Intel specifications....

It's not clear to me that that is the correct interpretation. The specific VID (voltage ID) value programmed into each Core2Duo CPU (which has a theoretical range of 0.85V-1.5V for e2xxx, e4xxx, and e6xxx CPUs, and a range of 0.85V-1.3625V for e8xxx CPUs) is Intel's maximum *for that CPU*. The reason for the range is to allow for manufacturing variation. Thus, any voltage above the VID voltage is "exceeding Intel specifications" *for that specific CPU*. For relatively small voltage boosts (on the order of 0.1V), I wouldn't expect there to be much if any harm. However, if you had a CPU with a VID value of 0.85V, I certainly wouldn't run it at 1.5V, which would be over a 75% voltage boost! Let's also remember that the Intel "absolute maximum" voltage for e2/4/6xxx is 1.55V.
Here's what Intel says about "functional" limits (e.g. VID) and "absolute maximum" limits:
Quote:
At conditions outside functional operation condition limits, but within absolute
maximum and minimum ratings, neither functionality nor long-term reliability can be
expected. If a device is returned to conditions within functional operation limits after
having been subjected to conditions outside these limits, but within the absolute
maximum and minimum ratings, the device may be functional, but with its lifetime
degraded depending on exposure to conditions exceeding the functional operation
condition limits.
At conditions exceeding absolute maximum and minimum ratings, neither functionality
nor long-term reliability can be expected. Moreover, if a device is subjected to these
conditions for any length of time then, when returned to conditions within the
functional operating condition limits, it will either not function, or its reliability will be
severely degraded.

February 6, 2008 7:11:06 PM

Mondoman said:
It's not clear to me that that is the correct interpretation. The specific VID (voltage ID) value programmed into each Core2Duo CPU (which has a theoretical range of 0.85V-1.5V for e2xxx, e4xxx, and e6xxx CPUs, and a range of 0.85V-1.3625V for e8xxx CPUs) is Intel's maximum *for that CPU*. The reason for the range is to allow for manufacturing variation. Thus, any voltage above the VID voltage is "exceeding Intel specifications" *for that specific CPU*. For relatively small voltage boosts (on the order of 0.1V), I wouldn't expect there to be much if any harm. However, if you had a CPU with a VID value of 0.85V, I certainly wouldn't run it at 1.5V, which would be over a 75% voltage boost! Let's also remember that the Intel "absolute maximum" voltage for e2/4/6xxx is 1.55V.


I understand what Intel has to say, and I get what you say. And I had forgotten to account for the Penryn line, so I appreciate that.

I was told by someone much more knowledgable than myself about the same issue. Essentially when manufacturing processes are identical they just use the lowest VID possible to provide a stable chip. If you have a VID of 1.125 and run it at 1.350 there is are no potential issues that arise because since it does not exceed the Intel Max.

This is what I was taught, and I may be wrong... however I know people who have had Core2s running on 1.5v since they day they came out with no problems. Obviously any CPU will eventually have electromitigation - and increasing voltage will decrease the lifespan, however I have never had an issue with it (though it's more and more a potential issue as we shrink processes).
February 6, 2008 7:50:25 PM

In light of all this I think for the time being I will run at 3.2, which for me is stable at vcore 1.312. This is the upper end of the VID voltage range as set forth in the Intel spec sheet for the e4600.

But Mondoman, the VID voltage range is 1.162-1.312 according to the spec sheet. As I understand what you're saying, the max VID for my particular chip could be anywhere in that range. If this is true, isn't anyone running above the low end of the range at risk? (Although I understand that even if 1.162 is the "correct" VID for my chip, running at 1.312 is not excessive. Also, my Bios puts the VID at 1.2875. Is that somehow the accurate VID for my chip?

Thank you both for this educational discussion. It is very good of you to take the time to pass on your learning.
February 6, 2008 11:26:02 PM

cnumartyr said:
... Essentially when manufacturing processes are identical they just use the lowest VID possible to provide a stable chip.
Yes, but the fact that different chips have different lowest VIDs means that the features within the chips themselves *are not identical*, even though they were made on the same process. Thus, they may very well have different sensitivities to increased voltage.
cnumartyr said:
...
If you have a VID of 1.125 and run it at 1.350 there is are no potential issues that arise because since it does not exceed the Intel Max.
While I agree that problems are very unlikely in this specific case, I don't agree that there are "no potential issues", since 1.35V is (according to my reading) "...outside functional operation condition limits..." for this specific CPU and thus "...the device may be functional, but with its lifetime degraded depending on exposure to conditions ...". My main concern is that we not mislead OCers into thinking that boosting the voltage above the VID is sanctioned by Intel and doesn't void the CPU warranty.
I find it very hard to believe that a CPU with a VID of 0.85V can run indefinitely at 1.5V without any issues at all. Now, perhaps I'm reading too much into Intel's published info. 0.85V is just the lowest possible VID using VID bits 1-6; perhaps no such e2/4/6xxx parts exist in real life, and the true range of real parts has VIDs with a range something like 1.1V-1.4V. The fact that rh's e4600 has a spec with a narrow 1.162-1.312V VID range supports this hypothesis.
February 6, 2008 11:34:03 PM

red hook said:
... the VID voltage range is 1.162-1.312 according to the spec sheet. As I understand what you're saying, the max VID for my particular chip could be anywhere in that range. If this is true, isn't anyone running above the low end of the range at risk?

Not exactly; I would say that anyone running above the specific VID *for the particular chip* is "at risk", although for small voltage increases that "risk" will be minimal. My main point is just that running a higher voltage than the VID is not "within Intel's specs" or "approved" by Intel, even if that voltage is within the spec range of VIDs for that chip model. It's not something I would personally worry about, but for some people keeping their warranty is important.


February 6, 2008 11:38:20 PM

Mondoman said:
Not exactly; I would say that anyone running above the specific VID *for the particular chip* is "at risk", although for small voltage increases that "risk" will be minimal. My main point is just that running a higher voltage than the VID is not "within Intel's specs" or "approved" by Intel, even if that voltage is within the spec range of VIDs for that chip model. It's not something I would personally worry about, but for some people keeping their warranty is important.

Can Intel tell if you have exceeded the specified VID?
a b à CPUs
a b K Overclocking
February 7, 2008 6:01:11 AM

Guys, to help lend a different perspective to this topic, the following Tom's article from this past July 17th shows that an OEM is willing to stake their reputation on high Vcore, and is apparently willing to stand behind it:

Overclocking Marathon Day 1 - A Screamin' Dell - http://www.tomshardware.com/2007/07/17/system_builder_m...

Please review paragraph 6 - "Overclock Settings" -


"Dell doesn't allow its end users to modify voltage levels, even on its overclocking machine. The system arrived with the CPU at a fixed 1.60 volts."


The Dell system in this article is running a liquid cooled QX6800. One can't help but wonder, what does this imply regarding VID, Vcore, thermal disipation efficiency, longevity and Intel warranty issues? Additionally, as no one has yet mentioned, VID is used by BIOS to establish minimum bootstrap Vcore.

Comp :sol: 
a b à CPUs
a b K Overclocking
February 8, 2008 8:35:55 AM

red hook, cnumartyr, Mondoman and Evilonigiri, I'm very interested in reading your thoughts on this article. Does anyone care to comment?

Comp :sol: 
February 8, 2008 11:23:42 AM

CompuTronix said:
red hook, cnumartyr, Mondoman and Evilonigiri, I'm very interested in reading your thoughts on this article. Does anyone care to comment?

Comp :sol: 


My personal view is that either Dell engineers are confident enough that they will eat the cost if the CPU fails or there are just no issues in their eyes.

I want to say it validate's us "overclockers" but at the same time, I don't know. I really think Intel has max voltages for a reason and as a manufacturer are "required" to state the "increasing voltage beyond VID can cause severe damage" line.

Either way I don't think electromitigtation will set in on the 65nm C2Q's quickly due to a voltage increase to 1.5 or even 1.6 vCore.
February 8, 2008 11:48:41 AM

All: please keep in mind that I'm here to give a newbie's perspective, and adjust yopur expectations accordingly.

1. I had hoped to perform a gotcha and point out that Dell's price for an extended warranty on this system was bocu $, but it is not so.

2. Of course Dell has for some time been trying to capture a bigger share of the high-end market, and this could be an overstep to accomplish that purpose.

3. Intel's spec sheet for the QX 6800 show a VID voltage range of 0.85-1.5. Perhaps they interpret Intel's advice as meaning that the voltage range applies to a particular processor type, rather than an individual CPU, i.e., 1.5 is OK for any 6800. In that case, 1.6 is only modestly above spec.
February 8, 2008 9:20:35 PM

CompuTronix said:
red hook, cnumartyr, Mondoman and Evilonigiri, I'm very interested in reading your thoughts on this article. Does anyone care to comment?

Comp :sol: 

Assuming that the 1.6V is the VID, the real voltage would probably be within Intel's specs.

But anyways, they probably did a statistical analysis, and found out that 1.6V covers 99% (my guess) of the cpus at that speed. That way, all they need to do is set it to 1.6V and be done with it, without doing any stability test. This effectively things speeds up, so they can make more money...
!