Inconsistant V-core at stock

afflictionau

Commendable
Mar 4, 2016
11
0
1,510
Hey everyone,

I am planning to overclock my CPU and want to get it high enough to support my 3200MHz RAM but I am very new to overclocking and was reading through some of the stickies on this forum. I was doing a benchmark baseline for my CPU at stock when I noticed that CPUID's Hardware monitor was recording a very high Vcore [~1.45v] so I checked the hardware monitor that came with my motherboard utilities and noticed it was significantly lower.

If the CPUID one is correct then the CPU at stock is at a higher Vcore than the recommended maximum of this CPU of 1.35v. I was wondering if anyone could help me work out if this is just a problem with the software or if it actually is running higher than it should be.

My Build:
Intel i5 6600K
Corsair H60 SE
Gigabyte Z170X - Gaming 5
Corsair Vengeance LPX 16GB (2*8GB) @ 3200Mhz (OC) [Currently at 2133Mhz]
Antec High Current Gamer (900W)
XFX Radeon R9 380 (4GB)

The Vcore recordings:


Hopefully this can be resolved
 
The new Skylake CPU has moved the voltage regulator from the CPU to the motherboard. Intel has not specified a particular recommended Vcore of these unclocked chips at stock. Motherboard manufacturers have thus decided on a value for themselves. In order to be CERTAIN with the new Skylake chip, they have often specified a rather high Base Vcore which the CPU can increase under demand. Many users report very high Vcores on some motherboard.

The first thing you could do is to use the UEFI/BIOS to undervolt your CPU.

Do you plan to overclock it? If so we can combine the two together, or you can undervolt now and overclock later.
 
Hey DonkeyOatie, Yeah I do plan to overclock, I'm just a little confused is all because one of the stickies on this forum "Intel temperature guide" states that the i series 6th gen max recommended Vcore was 1.35v.
However I had noted a few posts about the Vcore for this generation being unusually high. So I'd believe what you are saying.

As I am completely new to this and know that there is always more to learn and read about overclocking any help with the process would be greatly appreciated.

Are there any potential BIOS settings or otherwise that would be boosting the Vcore past what is actually set to in the BIOS? From memory it's currently set to something around 1.1-1.2
 
I don't think so. I think you are seeing the default of the board. My uncertainty comes from not having your system in front to me to look at and test :)

The first thing I would do is go into BIOS/UEFI (which I'm going to refer to as BIOS) and set the Vcore manually to 1.200V and then run some tests and stresses on your system. Do you know how to do this? If not start reading and researching :) I'll be happy to answer questions, but you need to understand this for yourself.

What are you using to test and stress your system (or plan to use)? I can recommend a bunch of good free stuff.
 
Yeah okay, I do understand that.
Yeah I can do that. As a sudden thought, could turbo mode be effecting the Vcore?

I would normally use Prime95 and just use Small FFT's as this change will only be affecting CPU so I don't see the need to include RAM testing :)
What would you suggest?
 
Yes, Turbo is affecting Vcore, but so it should. The 'problem' is that the Vcore is just too $%^$#@# high!

If you use Prime95 (and I do) use version 26.6. Anything more modern that that is dangerous to these processors.

OCCT and ASUS RealBench both provide good real benchmarks and stress tests. In particular OCCT will cause failures on seconds, that Prime97 take minutes to find (but things that are stable under OCCT may fail after hours of Prime95) you need to be using something to monitor your temperatures, OCCT will do that or HWMonitor or similar, which you seem to be using already.
 
At this point I'm really wishing I'd bought the MSI Mboard I was looking at, it had check points built into the board.

So now that I have changed the Vcore in BIOS from Auto (it displayed 1.125V) to a manual 1.2V
CPUID HW monitor is reporting 1.5V meanwhile Gigabyte's SIV is reporting 1.14V.

I'm actually getting fairly concerned by this point. Why is the software that most people believe is most reliable giving me values that are so HIGH?

I am noticing though that whether the computer is barely doing anything or whether I'm running Prime95, the Vcore is staying fairly consistent... Unfortunately the consistent value is 1.5-1.6V. Even though it should be dropping when not working so hard.
 
From BIOS I went into advanced voltage settings and then to CPU core voltage control and changed CPU Vcore from auto to 1.2v

I had left the CPU Vcore Loadline Calibration as auto.(under advanced power settings) but changing it to normal as I expected has definitely removed some of the variation however it is still very high compared to what it should be.
Again we have CPUID's HW Monitor at ~1.57v and SIV Hardware monitor at ~1.18V
Disregarding the variation for the moment, is there actually a clear reason why these two programs could be so different? Gigabyte's SIV always seems to be fairly close to the set Vcore where as CPUID is no where near what it is set to.
 
Don't know why, I use ASRock boards mostly, although Gigabyte makes a good product, but Skylake is so new, I still don't have enough specific experience (and comparatively few do) . I'll be doing some work on this in the summer, during my vacation.

What are your temperatures like? 1.5+ V should be very hot.
 
While doing a Prime95 run the max temperature I got was only 52C, I am using water cooling but it's not exactly a massive cooler, it was a cheap-ish 120mm Radiator "Corsair H60 SE"
This is one of the reasons I'm starting to suspect that CPUID is lying to me. If I check the Vcore in the bios' monitoring section it's listed to be very similar to the hardware monitor I have from Gigabyte (The SIV one)
As shown here:
 
Well I guess I should check some other HW monitors to make sure but yeah it's definitely strange. I was running small FFT's in prime95 for about 10 minutes or so, it's not the most major test for heat, I was mainly just checking the Vcore. That being said, today is a surprisingly cold day so the tests will be lower temperatures even if I did Large FFT's

I guess in summary, are there any other HW Monitors you would suggest me checking?
 
I've had it running for 15minutes now, max CPU temp reached was 45C, I did drop the Vcore back down to 1.125V though, just becasue it being labeled as over 1.5v was off putting.

OCCT is listing the same Vcore as HW Monitor... Is there any way to definitively check the Vcore? This variation between software is really concerning me, I'm not confident with overclocking until I can resolve this.

So the Vcore in SIV didn't even get to 1.125v meanwhile...