Some questions about reading core voltage on MSI Z87I AC with I5-4690k

[Taavn]

When using Hwinfo64 with C-states on, C1E support on, and manual voltage set to 1.25V in my BIOS, under the Vcore reading in Hwinfo64 I'm seeing fluctuating Vcore as if manual voltage was turned off.  This doesn't seem normal.  My understanding is that it should stay at the voltage I set it at regardless of power settings.  CPU-Z shows a fixed core voltage under these same settings.

If I turn C-States and C1E support off, but leave on manual voltage, the Vcore stays set at 1.264V in Hwinfo64.  In Hwmonitor the Vcore seems bugged and stays at 0.888V no matter what, but the VIN7 value always corresponds to the Vcore value in Hwinfo.  CPU-Z shows a core voltage of 1.252V.  

I noticed that when stress testing using Prime 95 26.6 I see the Vcore value in Hwinfo64 and the VIN7 in Hwmonitor jump to 1.28V very quickly, while CPU-Z shows 1.252V.  I shouldn't be seeing these kind of voltage increases because I'm using the older version of Prime 95 that isn't using AVX instructions right?

So, which readings are correct and which aren't?

 

[Martin]

Those CPUs have a Fully Integrated Voltage Regulator (FIVR), which means the voltage regulation is completely integrated in the CPU and I'm not sure how well the logic used on particular mainboard is effective in monitoring the internal Vcore.
Have you tried to compare the values HWiNFO reports with GIGABYTE EasyTune tool ?

 

[Taavn]

Gigabyte EasyTune on a MSI motherboard? Do you mean MSI command center or are you specifically recommending EasyTune for a reason?

I do notice that in the Bios itself the Vcore readings tend to be around .01-.02V higher than what I manually set it, but only when overclocking.

 

[Martin]

Ah sorry - I meant MSI Command Center. I mixed that because I have been recently dealing with GIGABYTE boards too
So it would be interesting to see the Vcore correlation between HWiNFO and MSI Command Center (side-by-side).

 

[Taavn]

Hwinfo64 and MSI Command center seem to match up. I'm guessing Hwinfo is the correct value. Although I'm still wondering why the voltage drops when set to override in the bios, when I have the cstates and c1e turned on. I thought the voltage override is supposed to, well override that.

 

[Martin]

As for the accuracy of Vcore monitoring, I think MSI should know best.
You might also check in the main HWiNFO window under CPU / Overclocking the exact parameters of "IA Overclocking" - if the manual voltage you set in BIOS is also properly set in the CPU.

 

[Taavn]

Under IA Overclocking I see a value called "target voltage" and it's set to the manual voltage I put in the bios. Is that what you're referring to?

 

[Martin]

Yes, that's it. So it's properly set.
I'm afraid, I can't give you an answer about what you're seeing on Vcore, I suggest to check with MSI.