Idle voltages: Vcore vs VR VOUT

napata · Aug 7, 2021

Everyone says that VR VOUT is the most accurate voltage reading of the die of the CPU, but when I look at my idle voltages then VR VOUT just doesn't seem to make sense. When my multiplier drops to 8x my Vcore reading drops but the VR VOUT reading never drops at all. It only ever drops during load from Vdroop.

Here's what the voltage reading looks like in a semi-idle state:

Can someone help me understand this please? It makes sense that voltages drop at idle so I'd would think the Vcore is the accurate one here?

Also something else I don't understand: my VR VOUT is always a couple of mV higher in idle than the voltage set in the BIOS. This happens at every LLC setting. I thought it would be lower due to Vdroop with a larger difference at a lower LLC? The actual BIOS reading is also higher for some reason as if there's a hidden offset to the set voltage.

The above readings are with a 1.372 adaptive voltage set in the BIOS voltage with the lowest LLC, though the BIOS reports the voltage around 1.383. I'm not sure which sensor the BIOS uses here. As you can see the VR VOUT also goes as high as 1.384 occasionally, but mostly stays around the 1.375 range. Not sure if it matters but I'm on a Z590 Gigabyte mobo with RL.

Nikado7 · Aug 7, 2021

I've noticed the same thing. Sub'ing.

AndrejB · Aug 7, 2021

Search falkentyne's posts around here and reddit. He explained it in detail.

Vrout (at least on z390 auros master is the closest to what you would see with a oscilloscope).

GeneO · Aug 7, 2021

I saw this on my Z390 Aorus Master as well and a satisfactory explanation never surfaced. It is affected by whether you have speedshift enabled or not. With speedshift disabled (using speedstep, aka EIST, instead) , VR VOUT more closely tracks Vcore.

acoustic · Aug 7, 2021

GeneO said:
I saw this on my Z390 Aorus Master as well and an adequate explanation never surfaced. It is affected by whether you have speedshift enabled or not. With speedshift disabled (using speedstep, aka EIST, instead) , VR VOUT more closely tracks Vcore.

I can't say I see this with my 10900K on an MSI Z490 ACE. VR VOUT reads the same whether I'm using Speed Shift or EIST. For 24/7 purposes, I run EIST disabled with Speed Shift enabled

GeneO · Aug 7, 2021

What I was talking about on the z390 master.

Starting here:

(Gigabyte Z390 AORUS Owners Thread)

This thread is still under construction. Please be patient and thank you for understanding. This thread is for the Z390 Aorus motherboard owners or future owners. We would like to be able to make this the go to for info, help, complaints, and suggestions about the board. Lets make this thread...

www.overclock.net

An eventually ending here:

(Gigabyte Z390 AORUS Owners Thread)

This thread is still under construction. Please be patient and thank you for understanding. This thread is for the Z390 Aorus motherboard owners or future owners. We would like to be able to make this the go to for info, help, complaints, and suggestions about the board. Lets make this thread...

www.overclock.net

GeneO · Aug 7, 2021

acoustic said:
I can't say I see this with my 10900K on an MSI Z490 ACE. VR VOUT reads the same whether I'm using Speed Shift or EIST. For 24/7 purposes, I run EIST disabled with Speed Shift enabled

I think it is a Gigabyte thing.

Falkentyne · Aug 8, 2021

napata said:
Everyone says that VR VOUT is the most accurate voltage reading of the die of the CPU, but when I look at my idle voltages then VR VOUT just doesn't seem to make sense. When my multiplier drops to 8x my Vcore reading drops but the VR VOUT reading never drops at all. It only ever drops during load from Vdroop.

Here's what the voltage reading looks like in a semi-idle state:
View attachment 2520211

Can someone help me understand this please? It makes sense that voltages drop at idle so I'd would think the Vcore is the accurate one here?

Also something else I don't understand: my VR VOUT is always a couple of mV higher in idle than the voltage set in the BIOS. This happens at every LLC setting. I thought it would be lower due to Vdroop with a larger difference at a lower LLC? The actual BIOS reading is also higher for some reason as if there's a hidden offset to the set voltage.

The above readings are with a 1.372 adaptive voltage set in the BIOS voltage with the lowest LLC, though the BIOS reports the voltage around 1.383. I'm not sure which sensor the BIOS uses here. As you can see the VR VOUT also goes as high as 1.384 occasionally, but mostly stays around the 1.375 range. Not sure if it matters but I'm on a Z590 Gigabyte mobo with RL.

I don't remember. VR VOUT showing something like 1.25v when Vcore (Super I/O) is showing 0.70v is something to do with power saving turning off part of the CPU/VRM area, after the voltage coming from the IR 35201 is already read from the VRM signal lines. I know nothing more about this but this was discussed in the Z390 thread some time back.

napata · Aug 16, 2021

GeneO said:
What I was talking about on the z390 master.

Starting here:

(Gigabyte Z390 AORUS Owners Thread)

This thread is still under construction. Please be patient and thank you for understanding. This thread is for the Z390 Aorus motherboard owners or future owners. We would like to be able to make this the go to for info, help, complaints, and suggestions about the board. Lets make this thread...

www.overclock.net

An eventually ending here:

(Gigabyte Z390 AORUS Owners Thread)

This thread is still under construction. Please be patient and thank you for understanding. This thread is for the Z390 Aorus motherboard owners or future owners. We would like to be able to make this the go to for info, help, complaints, and suggestions about the board. Lets make this thread...

www.overclock.net

Thanks for finding these posts. It's like you said disabling Speedshft "fixes" this.

Falkentyne said:
I don't remember. VR VOUT showing something like 1.25v when Vcore (Super I/O) is showing 0.70v is something to do with power saving turning off part of the CPU/VRM area, after the voltage coming from the IR 35201 is already read from the VRM signal lines. I know nothing more about this but this was discussed in the Z390 thread some time back.

It does seem to be related to the power saving settings but I still find it a weird, seemingly contradictory reading. Seems like the power saving features cause some issues with the sensors in general. The 0°C bug also seems to be related to power saving settings.

Or it could be a Gigabyte thing. I'm also getting 0 amps on IOUT most of the time while idle or browsing and it's also much lower than what you'd expect during load. This seems to affect CPU package power as well. Gigabyte doing Gigabyte things.

GeneO · Aug 16, 2021

napata said:
Thanks for finding these posts. It's like you said disabling Speedshft "fixes" this.

It does seem to be related to the power saving settings but I still find it a weird, seemingly contradictory reading. Seems like the power saving features cause some issues with the sensors in general. The 0°C bug also seems to be related to power saving settings.

Or it could be a Gigabyte thing. I'm also getting 0 amps on IOUT most of the time while idle or browsing and it's also much lower than what you'd expect during load. This seems to affect CPU package power as well. Gigabyte doing Gigabyte things.

On the Gigabyte, IIRC, there were two VRVOUT/IOUT. One for the CPU VRM, and one I believe for the memory. You may be looking at the IOUT for the second.

Intrud3r · Aug 16, 2021

This is my gigabyte with speedshift enabled and eist disabled (c-states disabled)

Voltage = VR VOUT
Amps = Current IOUT

(need to remake screenshot, wasn't a good comparison, too much crap running in the background)

after reboot, standard background programs running, about 3 min idle:

Intrud3r · Aug 16, 2021

This is with speedshift disabled and eist enabled (c-states disabled)

GeneO · Aug 16, 2021

Yeah that is what I was seeing with my Z390 Aurous master. It is almost as if SVID is disabled with speedshift on these boards. I don't see this on my Asus (with die sense vcore instead of VRVOUT). I don't know if it is incorrect reporting of VRVOUT or actually the VRVOUT is correct.

BINARYGOD · Jan 12, 2024

I necro this only as update to let it be known that these issues persist across multiple BIOS on a z690 Aorus Ultra (13700K, 13900K, 14900K) and across two BIOS on z790 Aorus Master (2 different Masters, but on these two, I only used the 14900K).

I am considering running away to the ASRock Z790 NOVA - but I have had my fill of reseating processors, RMA-ing, and returning to Amazon, and of such these past 12 months.

Idle voltages: Vcore vs VR VOUT

napata

Nikado7

AndrejB

GeneO

acoustic

GeneO

(Gigabyte Z390 AORUS Owners Thread)

(Gigabyte Z390 AORUS Owners Thread)

GeneO

Falkentyne

napata

(Gigabyte Z390 AORUS Owners Thread)

(Gigabyte Z390 AORUS Owners Thread)

GeneO

Intrud3r

Intrud3r

GeneO

BINARYGOD

Top Contributors this Month

Recommended Communities