Which VCore?

[Djang0 1]

I set VCore manually in BIOS but get the following measurements (motherboard is a Gigabyte EP45-UD3LR):

BIOS = 1.356V

CPU-Z = 1.328V

HWMonitor = 1.33V (1.28V min, 1.33V max)

 

Why is there a difference?

Which value should I trust?

 

The recommended VCore max for my Q9650 is 1.3625V which is the next notch on the list of VCore's in BIOS and I don't want to go over this value (yes I know that some people happily run their Q2Q at 1.4somethingV but I don't!). Of course if the real VCore is 1.33V when my BIOS says 1.35V then perhaps I can go a few notches above 1.3625V in the list. Thanks in advance for your replies.

[dazz 69]

Max full load Vcore is what matters.

 

It's normal to have a lower Vcore once booted into the OS compared to what you set in your BIOS. It's called Vdrop.

Then there may be a further drop as load is applied to the CPU. That's called Vdroop, and can be mitigated using Load Line Calibration (LLC) . Some mobo vendors call LLC different things like MSI (Vdroop control)

 

It's also normal to have small fluctuations in Vcore like those 1.28 - 1.33V you see there.

The difference between CPU-Z & HWMonitor is simply a question of precision. CPU-Z uses 3 decimal digits vs 2 of HWMonitor, which simply rounds 1.328 up to 1.33V

 

Furthermore, with C-States enabled, you may see your Vcore drop as low as 0.9 -1V when the CPU is idling

[Djang0 1]

Thanks for your reply. For overclocking I always disable the C-states although after I've have found a stable overclock I try to tun EIST. As for LLC it's not recommended for Q2Q's apparently because of dangerous voltage spikes, not sure about other processors though.

 

To be honest I'm a bit confused about how VCore relates to VID even though an explanation of VID is given in that article! The Intel spec says "VID Voltage Range" = 0.8500V - 1.3625V, so perhaps I'm misunderstanding VID max as VCore max? In that case I should be paying more attention to VID. E.g. when:

VCore (BIOS) =1.3500V

VCore (CPU-Z) = 1.312V

VID (CoreTemp) = 1.2750V

(Sorry to change the figures from before I'm actually overclocking as I write this!)

The difference between VID max = 1.3625V and the above example of VID = 1.2750V gives me far more room to play with... perhaps those folks at VCore = 1.4something aren't that crazy! Does this make sense?

[dazz 69]

Focus on full load Vcore readings in CPU-Z and forget about VID. VID is only meaningful at stock speeds, and even then, you'll only find out how good your chip is for OC once you start pushing it.

1.31V looks good to me

[Djang0 1]

Focus on full load Vcore readings in CPU-Z and forget about VID

Yeah I see what you mean, I'm confusing mysefl! Just for reference my VID is 1.2750V. What I appreciate now is there is no Intel recommended Vcore max (1.3625 is just the upper VID - if you're unlucky!).

 

So can I push VCore as far as my temps allow? Right now I'm just aiming for 3.99GHz but as the Q9650 is older hardware I'd like to push it higher.

[TechguyMaxC 516]

Disable vdroop if you want to see your read vcore values get closer to what you have set them to in the BIOS. Some motherboards call this feature Load Line Calibration (LLC) rather than vdroop.

[Djang0 1]

Disable vdroop

Please see my previous post re LLC.

 

OK so I did some browsing of overclocking forums and it appears that for a Q9650 people are throwing figures around of Vcore max (apparently the lingo is "Vccmax") of 1.4V to 1.45V for 24/7 use and that temperatures will likely be the limiting factor. I'd really like to know where those figures come from?

 

I'm tempted to push towards 1.4V unless the gains become foolishly deminishing. On the other hand my 775 board is only mid range so perhaps I should be happy if I can get it stable at 3.99GHz and run my memory at stock (albeit at 333MHz frequency latch rather than 400MHz)?

 

By the way I should mention that I went from struggling to keep temps < 70C at 3.85GHz to pushing towards 4GHz with temps < about 65C by simply moving my PC case from under my desk to by my window where there is better ventilation and a draft! I can keep temps even lower if I open other windows in my house which increases the draft. And to think I was considering upgrading to a £80+ water cooling kit - no need to now!

[TechguyMaxC 516]

I had a C2Q 9550 overclocked to 4GHz within 5 minutes of getting it out of the box. Ran about 1.44V, IIRC and never had stability or temperature concerns over the course of about 2 years. Also ran LLC to minimize vdroop as much as possible. Advice you'll find online WRT overclocking and the proper or safe application of voltage tends to err on the side of caution, often by far more than necessary.

[Djang0 1]

So I thought I'd give my Q9650 system one more round of overclocking before I build my next i7-4790K system. Results below - still testing stability and not so happy with my temps under stress testing.

 

Q9650 OC:

CPU @ 4GHz  = 9 x 445MHz

DDR2-800 @ 990MHz = 445 x 2 (f-latch = 333MHz / 1.8v)

Vcore idle:

BIOS setting      = 1.38750v (???)

BIOS measured = 1.348v - 1.364v

CPU-Z               = 1.328v - 1.344v

HW Monitor       = 1.328v - 1.344v

EasyTune6        = 1.344v (N.B. OCing from BIOS)

Vcore load:

CPU-Z = HW Monitor = Easytune6 = 1.296v

Temperatures:

Stress test = 60C - 65C but core0 up to 70C

FSX           = up to 60C

 

Q) Once again which Vcore should I use to guide how high I set the BIOS value?