Max CPU Vcore in bios exceeds Intel Specs
I have a P35-DS3L and a E2200 processor. Can someone educate me a little on core voltage. Intel's data sheet for the E2200 says the VID voltage range is 1.162V-1.312V. However in my bios the Max CPU Vcore is listed as 1.325 which is higher than what Intel's range is for this processor. And I have seen some people's data for their overclock that shows some have run the core voltage as high as 1.5. Can someone explain this?
But I am not talking about being overclocked. With no overclocking, my bios reports the Max CPU Vcore as 1.325. I am not talking about the actual Vcore which you can change. By default both are the same but why would my bios report the max as higher than Intel specs. Also, can't running the core voltage higher than Intel specs damage the processor? If so, why do so many people do it?
The processor's Voltage ID (VID) is read by BIOS, just before the beep code, during the motherboard's initial powerup sequence, to identify and configure the proper voltage required to initialize a stable CPU boot. Vcore BIOS, to which you're referring, is not VID. What you see in BIOS are 2 values; Vcore "setting", usually located under "advanced", and Vcore "applied", usually located under "power".
This means that the "applied" Vcore is a response to the Vcore "setting" you've selected, whether it's auto or manual. The actual VID, or boot voltage is "masked", and is typically not shown in BIOS. However, I had an MSI motherboard (Socket 939 AMD) a few years ago which had BIOS that displayed boot voltage (VID) and Vcore applied, but I haven't seen another BIOS with this feature before or since.
VID can be read in utilities such as Real Temp 2.7, and is most useful as an indicator of a processor's overclocking potential. The lower the VID, the less Vcore is needed to boot the processor, or to reach a higher overclock; thus it is said that you have a "better" chip. Intel typically releases the same processor variant with the same stepping with different VID's, because some simply test better than others. Given consistent purity in silicon, as the fabrication techniques mature over time for a particular processor variant, VID's typically become lower.
The difference between Vcore idle and Vcore load (Vdroop) is simply the inherent electronic characteristics of the voltage regulators responding to an increase in load. Vcore is created on the motherboard by the voltage regulators, which are supplied with 12 volts directly from the PSU via the 4 or 8 pin connector near the CPU socket. When CPU workload increases, current (amperage) increases, which increases heat and resistance, and in turn causes Vcore to decrease slightly, or "droop". The better the motherboard design, such as 8 phase regulation, the less Vdroop. Some motherboards include a Vdroop compensation feature in BIOS.
When Vcore is manually increased for overclocking purposes, it typically exceeds VID. When auto Vcore is selected, VID ensures that Vcore won't be too low to successfully boot. This is why Vcore BIOS is almost always higher than when in Windows at idle with CPU-Z, which is a point that so many users find confusing, and don't understand. Regardless, the following is typical and expected:
Vcore BIOS setting = highest value
Vcore BIOS applied = higher value
Vcore Idle (CPU-Z) = lower value
Vcore Load (CPU-Z) = lowest value
Other minor variables include Analog to Digital (A to D) conversions, and Digital to Analog (D to A) conversions, which contribute to small inaccuracies between the Vcore being displayed both in BIOS and Windows, and the Vcore actually being applied to the processor (Vdrop). If Vcore at idle is slightly over spec, it doesn't present a problem for the processor in terms of electromigration or ablation, since there is little current flow, which results in relatively little heat being produced.
Vcore load in CPU-Z while running Small FFT's is lowest due to Vdroop, and is the most critical value, which is relative to the upper value shown as "VID Voltage Range" in Intel's Processor Spec Finder. Since heat is dissipated as a product of voltage and current, which is Watts, if CPU temperature and Core temperatures can be maintained below thermal spec by using high-end cooling, and Vcore load doesn't exceed specs, then you won't compromise processor longevity.
Concerning the "VID Voltage Range" for your E2200 M0, Intel shows conflicting specifications for the M0 stepping processors as follows:
E4700 (shown as G0 stepping) typo?
E4600 ??? typo?
E2200 ??? typo?
If you closely examine Intel's Processor Spec Finder - http://processorfinder.intel.com/Default.aspx - you can readily identify inconsistencies, which they're not too quick to correct. All Quad and Duo G0 variants are 1.5 Max, so the E4700 is listed incorrectly. All E2xxx variants are 1.5 Max, regrdless of stepping (G0, L2, M0), so your E2200 is listed incorrectly, as is the E4600.
Intel's "VID Voltage Range" as shown in the Proccesor Spec Finder can be quite confusing, since this rather poorly defined specification combines 2 parameters; "VID" and "Voltage Range". Using my Q6600 as an example, the first "Voltage Range" parameter is minimum Vcore run, which is the low value, and the second "Voltage Range" parameter is maximum Vcore run, which is the high value, as you can see below:
Intel's "VID Voltage Range" for my Q6600 G0 is 0.85 to 1.5, yet in the following example, "VID" is actually comprised of 2 more values, which are those used for Vcore boot. Real Temp 2.7 reads them individually as "Core VID Min and Max", but are not defined or shown seperately in Intel's Processor Spec Finder:
The higher value is again the most critical, since it indicates a processor's overclocking potential. Q6600 G0's have been released with VID's from 1.2 to 1.325, so my patricular VID at 1.3000 isn't the best, but it's good enough to run at 3.6 Ghz Prime95 small FFT's stabe for 12+ hours. Temperatures and Vcore Load are shown below in my signature. Q6600 G0's with the lowest VID's can reach 4.0 Ghz with high-end air cooling.
Intel sometimes doesn't use the common sense appraoch when writing specifications, so in these instances, we must refer to similar processor variants and revert to more resourceful means to make sense of their specification, such as researching hundreds of pages of Intel documents. In simple terms, however, I think that 2 seperate specifications would more clearly reflect processor voltages definitions such as:
VID: 1.1625 to 1.3
Voltage Range: 0.85 to 1.5
I'm quite certain that you're E2200 uses the same 1.5 Max as it's all M0 cousins, so feel free to overclock your processor and increase Vcore within the guidelines I've described in this post. If you still have your retail box, then check the sticker for voltage specs unique to your particular processor.
Concerning temperatures, understand that the "Thermal Specification" shown in Intel's Processor Spec Finder refers to CPU temperature, and not Core temperatures, which is another point that so many users find confusing, and don't understand. From my Temperature Guide linked below in my signature:
"Section 6: Scale ...
... as Tcase Max will be exceeded before Tjunction Max is reached, Tcase Max is always the limiting thermal specification...
Scale 1: Duo
E7200: Tcase Max 74c, Stepping M0,TDP 65W, Idle 8W
E4700: Tcase Max 73c, Stepping G0, TDP 65W, Idle 8W
E4x00: Tcase Max 73c, Stepping M0, TDP 65W, Idle 8W
E2xx0: Tcase Max 73c, Stepping M0, TDP 65W, Idle 8W
E8x90: Tcase Max 72c, Stepping C0, TDP 65W, Idle 8W
E8x00: Tcase Max 72c, Stepping C0, TDP 65W, Idle 8W
E6x50: Tcase Max 72c, Stepping G0, TDP 65W, Idle 8W
E6540: Tcase Max 72c, Stepping G0, TDP 65W, Idle 8W
Tcase = CPU temperature
Tjunction = Core temperatures
If you'd like to learn more about how CPU temperature and Core temperatures relate to one another, and if you'd like to learn how to calibrate these temperatures within 3c using SpeedFan, then check out my Temperature Guide.
I hope this explanation of VID, M0 stepping and temperatures answer all your questions,
Thank you so much CompuTronix for the education. I know a lot more now. I did still have my CPU box and it says 1.35V max.
One more question, is the goal to get the lowest voltage the the system will boot and run without errors or is the goal to get the highest voltage (closest to the max) the system will run at an not exceed heat specs?
Thanks again for the education.
Interesting, does the new Real Temp list the standard VID. (The way high one of 1.30) And then the minimum one... is the minimum the EIST stepped down version of your VCore? As in, is it listing what intel decided your x6 multi and reduced speed should now require for operational use? (A set value. Dependant on VID.)
That looks like an exact EIST VCore setting is why I ask. Like when it cuts the Multi and the speed, it can now cut the power as well, which older boards wouldn't do. Nice of it to list the value!
(or is it listing the minimum YOUR chip happened to go while Real Temp is open?)
And the extended VID range that intel lists covers the minimum Vdrop and droop range in spec. .85 V to 1.50 volts. So thats the operational voltage, isn't it? And the VID range would still be 1.2000 to 1.3250 (with 1.3500 added for some throw backs.)
I haven't tested the theory because I like to over clock, but supposedly the lowest VID chip should be able to function at the lowest voltage range. IE, a 1.2000 VID chip may work at stock while loaded at .85 volts?
I have only tested mine just under 1 volt, and I know that works!
(Didn't the Q6600 series used to be .95 to 1.50 when it was the B3, maybe? because I know the new series chips, q9xxx, have the same voltage range, supposedly. .85 to 1.3650 or whatever the exact upper value is, hehe!)
So many things, knowing such a simple value such as VID.
hell, even the basic info that VID provides, you'll note that so very few people even know. You get all the "stock" over clockers, hehe, saying they OC'd this or that on stock to x.xxx speed!
Most of them couldn't tell you to save their life, what their "stock" voltage was supposed to be, hehehe!
Good work, Compu! By any chance have you ever seen any technical documentation that lists the max intel spec vdrop and droop?
Everything I have seen points to .15 as being the worse drop/droop combo I have encountered on a Mobo without alterations.
(Though I can't figure out why sometimes I like to type alot, and dont mind, while other times I feel like hiding from the forums because I hear the same things over and over, hehehe!!)
HELP! Save me, CompuTronix!
You guys are both way over my head, but for what it's worth Real Temp is reporting my Core VID as Min = 1.1875 and Max = 1.3250. So know I have 2 questions.
1. Can I run core voltage safely up to 1.5 if necessary?
2. Is the goal to get the lowest voltage the the system will boot and run without errors or is the goal to get the highest voltage (closest to the max) the system will run at an not exceed heat specs?
The goal is to get both, kind of. You want to find a temp balance, along with VCore and CPU speed. (And memory speed, since its all tied in.)
Basically, you have the worse VID. 1.3250. The natural range is 1.2000, to 1.3250, in .0125 increments. IE, 1.2000 is the best! I have one! next up is 1.2125, 1.2250, 1.2375, on up to the worse, 1.3250.
There is an added VID of 1.3500, which they added for chips that totally failed their tests, so they make them the budget line, and all are 1.3500.
Obviously, the lower the starting voltage, the better the end result of EVERYTHING will be. Less heat, higher OC at less voltage, etc.
It is especially useful for NVidia chipsets! I can get 3.6 Ghz at 1.400 in the bios, to net 1.3 loaded. While a 1.3250 would need like 1.50 + in the bios to be stable due to the GIANT VDrop and droop NVidia is known for.
Vdrop = The initial voltage drop to the CPU when going from the BIOS to Idle in Windows.
VDroop = After above effect, but when loaded on prime 95 Small FFTs test on all cores. The voltage will again be reduced.
As over clockers, the worse this is, the worse everything is! (There are arguments to this.)
So once you establish what yours is, P35 DSL, those are like .075 or some junk, if I recall. I dislike Gigabyte! You'll know what you need to set to give you x.xxx while loaded on small ffts, as this is the most critical time for voltage.
Its easy to test!
Back to stock, disable speed step, EIST and C1E in the bios, set the CPU voltage to a manual 1.3250 and then re boot.
When in windows, wait a min for boot to complete, then look with CPUz and see what your idle voltage is in windows.
After that, just run small ffts torture test on all cores at the same time, wait 1 minute, and then watch to see what the lowest value is that CPUz now flickers to, or remains at.
List them here, and there you have it!
Your quad is more powerful than the specs make it look.
As long as you control the heat, there are plenty of speeds out there for you. 3.6 Ghz is obtained at around 1.44-1.45 volts while Loaded!
(List the voltages as Bios, Idle and Loaded values. Loaded is the most critical.) If I said I get 3.4 Ghz at 1.30 loaded. You know it means while running prime small ffts for 8 hours or more, it was stable at that lowest voltage.
But to answer your question, Yes, you can run 1.50 if you need to, and have the cooling for it. I regularly test up to 1.55 loaded. Which is the Spec Maximum voltage.
That can be around 1.60+ in a board like yours, bios value! hehe.
That would be the max I would use as far as voltage goes. Berserk testers use 1.9 volts and mad shiznit like that all the time with nitrogen cooling while setting huge test benchmarks!!
But to make it last years, no more than that, and thats with good water! 1.55 loaded. Maximum of 70c temps while testing for 8 hours. That will keep your daily temps down below 60c, and that should be fine for that chip. But more heat, or more voltage may start to wear at it.
I have no clue, because as far as i can tell, nothing hurts one! Hehe, but then again, others have blown them up!
My fingers hurt.
As lupi and I have pointed out, yes, it's OK to use Vcore Load 1.5 Max for your E2200. The "VID Voltage Range" shown in Intel's Processor Spec Finder used to be named something like "Core Voltage" which showed the same value as the retail box. Intel changed it last year to "VID Voltage Range", which reflected what overclocking experts had already determined; all 65 nanometer Core 2 processors will tolerate a maximum Vcore of 1.5 volts.
Also as lupi has pointed out, the goal is to reach the highest Prime95 stable overclock without exceeding 1.5 Vcore Load or 70c CPU temperature / 75c Core temperatures. Check out graysky's Overclocking Guide and my Temperature Guide at the top of this Forum.
I have learned alot from this old thread, but I have a problem. In my laptop I used to have a t6500 the VID
range was 1.000v to 1.250v. So now I have a q9000 in it, and the VID range for it is 1.050 to 1.175. Much
better right? With or without speedstep, in CPU-z or realtemp it idles at 1.050v. This is fine. However
under load without speedstep it will stay at 1.050v. not good. and with speedstep it only goes up to
1.112v. better but not good. The temps are always sub 70c (rates to100C). The ACPI tables are as
follows. 1.050v for 1.6ghz 1.112v for 2ghz and 1.225 for 2.3+ghz. The q9000 only goes to
2ghz. so im not getting that extra juice. The ACPI tables are the only things restricting me. Its driving me
crazy!! Could any of you please help. I would be forever in your debt.
PS: PC Wizard and some other applications say the VID is 1.167v sometimes.. are these just wrong?